Hi All, For a 2 level cache hierarchy with DDR3 (11-11-11) as the DRAM specs, the number of prefetch requests issued by the L2 prefetcher is high (very high in some cases) as compared to the number of prefetch requests responded by the DRAM controller. Any thoughts on it. The difference becomes almost negligible, when I use an L3 cache and turn on the prefetchers at the L2. In case of 3 level cache hierarchy, this equation holds true: pf_issued at the L2s = Pf hits at the L3 + Pf requests responded by the DRAM Here are the actual numbers for a 4 core setup with PARSEC benchmarks: blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 130935 # Number of read requests responded to by this memory blackscholes_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 148226 # number of hwpf issued canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 346384 # Number of read requests responded to by this memory canneal_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 354459 # number of hwpf issued dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 2937725 # Number of read requests responded to by this memory dedup_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 3251216 # number of hwpf issued facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 3443990 # Number of read requests responded to by this memory facesim_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 3639481 # number of hwpf issued ferret_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 2141141 # Number of read requests responded to by this memory ferret_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 12161061 # number of hwpf issued freqmine_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 115625 # Number of read requests responded to by this memory freqmine_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 168290 # number of hwpf issued rtview_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 590098 # Number of read requests responded to by this memory rtview_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 863745 # number of hwpf issued swaptions_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 20308 # Number of read requests responded to by this memory swaptions_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 16689699 # number of hwpf issued vips_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 3361213 # Number of read requests responded to by this memory vips_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 3542000 # number of hwpf issued x264_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 1060135 # Number of read requests responded to by this memory x264_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 1805457 # number of hwpf issued -- *thanks&regards* *BISWABANDAN* http://www.cse.iitm.ac.in/~biswa/ “We might fall down, but we will never lay down. We might not be the best, but we will beat the best! We might not be at the top, but we will rise.”

Hi Biswabandan, Is there any chance you could dig a bit deeper here? My suspicion is that the DRAM read queue is filling up and forcing packets to wait. The controller only counts reads when they are actually serviced. Perhaps the prefetcher counts issues prefetches as it creates the packets, even if they are not successfully sent? Check numRdRetry and the rdQLenPdf. Thanks, Andreas From: biswabandan panda via gem5-users <gem5-users(a)gem5.org<mailto:gem5-users(a)gem5.org>> Reply-To: biswabandan panda <biswa.uce(a)gmail.com<mailto:biswa.uce(a)gmail.com>>, gem5 users mailing list <gem5-users(a)gem5.org<mailto:gem5-users(a)gem5.org>> Date: Monday, 30 June 2014 12:19 To: gem5 users mailing list <gem5-users(a)gem5.org<mailto:gem5-users(a)gem5.org>> Subject: [gem5-users] DRAM Controller and Prefetchers Hi All, For a 2 level cache hierarchy with DDR3 (11-11-11) as the DRAM specs, the number of prefetch requests issued by the L2 prefetcher is high (very high in some cases) as compared to the number of prefetch requests responded by the DRAM controller. Any thoughts on it. The difference becomes almost negligible, when I use an L3 cache and turn on the prefetchers at the L2. In case of 3 level cache hierarchy, this equation holds true: pf_issued at the L2s = Pf hits at the L3 + Pf requests responded by the DRAM Here are the actual numbers for a 4 core setup with PARSEC benchmarks: blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 130935 # Number of read requests responded to by this memory blackscholes_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 148226 # number of hwpf issued canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 346384 # Number of read requests responded to by this memory canneal_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 354459 # number of hwpf issued dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 2937725 # Number of read requests responded to by this memory dedup_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 3251216 # number of hwpf issued facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 3443990 # Number of read requests responded to by this memory facesim_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 3639481 # number of hwpf issued ferret_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 2141141 # Number of read requests responded to by this memory ferret_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 12161061 # number of hwpf issued freqmine_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 115625 # Number of read requests responded to by this memory freqmine_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 168290 # number of hwpf issued rtview_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 590098 # Number of read requests responded to by this memory rtview_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 863745 # number of hwpf issued swaptions_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 20308 # Number of read requests responded to by this memory swaptions_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 16689699 # number of hwpf issued vips_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 3361213 # Number of read requests responded to by this memory vips_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 3542000 # number of hwpf issued x264_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher 1060135 # Number of read requests responded to by this memory x264_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued 1805457 # number of hwpf issued [https://ssl.gstatic.com/ui/v1/icons/mail/images/cleardot.gif] -- thanks&regards BISWABANDAN http://www.cse.iitm.ac.in/~biswa/<http://www.cse.iitm.ac.in/%7Ebiswa/> “We might fall down, but we will never lay down. We might not be the best, but we will beat the best! We might not be at the top, but we will rise.” -- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2557590 ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2548782

Hi All,
         For a 2 level cache hierarchy with DDR3 (11-11-11) as the DRAM

Hi Andreas, Thanks for the prompt response. Actually DRAM readQ is never filled up completely. Here are some of the stats: blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry 0 # Number of times read queue was full causing retry blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 29760 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 23193 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 18571 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 15509 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 12970 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 11623 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 10339 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 9681 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 3538 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 2324 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 1782 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 1758 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 1112 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 957 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 822 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 710 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 0 # What read queue length does an incoming req see blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry 0 # Number of times read queue was full causing retry canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 1820343 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 820262 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 214278 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 59606 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 47136 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 46901 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 45035 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 43390 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 38898 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 22687 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 11365 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 6793 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 3105 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 1925 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 1095 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 654 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 0 # What read queue length does an incoming req see canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry 0 # Number of times read queue was full causing retry dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 601147 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 466636 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 397007 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 320737 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 273422 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 235104 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 215016 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 204993 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 160861 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 104896 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 78910 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 84566 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 42656 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 32575 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 24300 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 18088 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 0 # What read queue length does an incoming req see dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry 0 # Number of times read queue was full causing retry facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 923820 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 570722 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 412947 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 336858 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 278383 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 242649 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 218409 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 205175 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 176151 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 159665 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 137869 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 129144 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 75569 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 52159 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 35059 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 24994 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 9 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 6 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 1 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 0 # What read queue length does an incoming req see facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 On Mon, Jun 30, 2014 at 6:29 PM, Andreas Hansson <Andreas.Hansson(a)arm.com> wrote: > Hi Biswabandan, > > Is there any chance you could dig a bit deeper here? > > My suspicion is that the DRAM read queue is filling up and forcing > packets to wait. The controller only counts reads when they are actually > serviced. Perhaps the prefetcher counts issues prefetches as it creates the > packets, even if they are not successfully sent? Check numRdRetry and the > rdQLenPdf. > > Thanks, > > Andreas > > From: biswabandan panda via gem5-users <gem5-users(a)gem5.org> > Reply-To: biswabandan panda <biswa.uce(a)gmail.com>, gem5 users mailing > list <gem5-users(a)gem5.org> > Date: Monday, 30 June 2014 12:19 > To: gem5 users mailing list <gem5-users(a)gem5.org> > Subject: [gem5-users] DRAM Controller and Prefetchers > > Hi All, > For a 2 level cache hierarchy with DDR3 (11-11-11) as the DRAM > specs, the number of prefetch requests issued by the L2 prefetcher is high > (very high in some cases) as compared to the number of prefetch requests > responded by the DRAM controller. Any thoughts on it. > > The difference becomes almost negligible, when I use an L3 cache and turn > on the prefetchers at the L2. In case of 3 level cache hierarchy, this > equation holds true: > pf_issued at the L2s = Pf hits at the L3 + Pf requests > responded by the DRAM > > Here are the actual numbers for a 4 core setup with PARSEC benchmarks: > > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 130935 # Number of read requests responded to by this > memory > blackscholes_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 148226 # number of hwpf issued > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 346384 # Number of read requests responded to by this > memory > canneal_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 354459 # number of hwpf issued > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 2937725 # Number of read requests responded to by > this memory > dedup_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 3251216 # number of hwpf issued > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 3443990 # Number of read requests responded to by > this memory > facesim_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 3639481 # number of hwpf issued > ferret_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 2141141 # Number of read requests responded to by > this memory > ferret_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 12161061 # number of hwpf issued > freqmine_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 115625 # Number of read requests responded to by this > memory > freqmine_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 168290 # number of hwpf issued > rtview_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 590098 # Number of read requests responded to by this > memory > rtview_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 863745 # number of hwpf issued > swaptions_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 20308 # Number of read requests responded to by this > memory > swaptions_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 16689699 # number of hwpf issued > vips_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 3361213 # Number of read requests responded to by > this memory > vips_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 3542000 # number of hwpf issued > x264_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher > 1060135 # Number of read requests responded to by > this memory > x264_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued > 1805457 # number of hwpf issued > > > > > > > > -- > > > *thanks&regards * > *BISWABANDAN* > http://www.cse.iitm.ac.in/~biswa/ > > “We might fall down, but we will never lay down. We might not be the best, > but we will beat the best! We might not be at the top, but we will rise.” > > > > -- IMPORTANT NOTICE: The contents of this email and any attachments are > confidential and may also be privileged. If you are not the intended > recipient, please notify the sender immediately and do not disclose the > contents to any other person, use it for any purpose, or store or copy the > information in any medium. Thank you. > > ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, > Registered in England & Wales, Company No: 2557590 > ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, > Registered in England & Wales, Company No: 2548782 > -- *thanks&regards* *BISWABANDAN* http://www.cse.iitm.ac.in/~biswa/ “We might fall down, but we will never lay down. We might not be the best, but we will beat the best! We might not be at the top, but we will rise.”

Hi All,
         For a 2 level cache hierarchy with DDR3 (11-11-11) as the

Just something to note that the prefetcher in previous versions of gem5 does not operate as it is supposed. A couple of patches were pushed in early this year to fix this issue, but I never checked to see if they really fixed the issue. The issue, if I remember right, is that prefetcher may generate multiple requests per cache line which is basically merged to one request in the DRAM read queue. I am not sure this directly relates to your problem... Thanks, Amin On Mon, Jun 30, 2014 at 7:50 AM, biswabandan panda via gem5-users < gem5-users(a)gem5.org> wrote: > Hi Andreas, > Thanks for the prompt response. Actually DRAM readQ > is never filled up completely. > > Here are some of the stats: > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry > 0 # Number of times read queue was full causing retry > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 > 29760 # What read queue length does an incoming req > see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 > 23193 # What read queue length does an incoming req > see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 > 18571 # What read queue length does an incoming req > see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 > 15509 # What read queue length does an incoming req > see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 > 12970 # What read queue length does an incoming req > see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 > 11623 # What read queue length does an incoming req > see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 > 10339 # What read queue length does an incoming req > see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 > 9681 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 > 3538 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 > 2324 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 > 1782 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 > 1758 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 > 1112 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 > 957 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 > 822 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 > 710 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 > 0 # What read queue length does an incoming req see > blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry > 0 # Number of times read queue was full causing retry > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 > 1820343 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 > 820262 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 > 214278 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 > 59606 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 > 47136 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 > 46901 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 > 45035 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 > 43390 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 > 38898 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 > 22687 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 > 11365 # What read queue length does an incoming req > see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 > 6793 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 > 3105 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 > 1925 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 > 1095 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 > 654 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 > 0 # What read queue length does an incoming req see > canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry > 0 # Number of times read queue was full causing retry > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 > 601147 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 > 466636 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 > 397007 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 > 320737 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 > 273422 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 > 235104 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 > 215016 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 > 204993 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 > 160861 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 > 104896 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 > 78910 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 > 84566 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 > 42656 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 > 32575 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 > 24300 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 > 18088 # What read queue length does an incoming req > see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 > 0 # What read queue length does an incoming req see > dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry > 0 # Number of times read queue was full causing retry > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 > 923820 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 > 570722 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 > 412947 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 > 336858 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 > 278383 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 > 242649 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 > 218409 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 > 205175 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 > 176151 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 > 159665 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 > 137869 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 > 129144 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 > 75569 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 > 52159 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 > 35059 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 > 24994 # What read queue length does an incoming req > see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 > 9 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 > 6 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 > 1 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 > 0 # What read queue length does an incoming req see > facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 > > > On Mon, Jun 30, 2014 at 6:29 PM, Andreas Hansson <Andreas.Hansson(a)arm.com> > wrote: > >> Hi Biswabandan, >> >> Is there any chance you could dig a bit deeper here? >> >> My suspicion is that the DRAM read queue is filling up and forcing >> packets to wait. The controller only counts reads when they are actually >> serviced. Perhaps the prefetcher counts issues prefetches as it creates the >> packets, even if they are not successfully sent? Check numRdRetry and the >> rdQLenPdf. >> >> Thanks, >> >> Andreas >> >> From: biswabandan panda via gem5-users <gem5-users(a)gem5.org> >> Reply-To: biswabandan panda <biswa.uce(a)gmail.com>, gem5 users mailing >> list <gem5-users(a)gem5.org> >> Date: Monday, 30 June 2014 12:19 >> To: gem5 users mailing list <gem5-users(a)gem5.org> >> Subject: [gem5-users] DRAM Controller and Prefetchers >> >> Hi All, >> For a 2 level cache hierarchy with DDR3 (11-11-11) as the >> DRAM specs, the number of prefetch requests issued by the L2 prefetcher is >> high (very high in some cases) as compared to the number of prefetch >> requests responded by the DRAM controller. Any thoughts on it. >> >> The difference becomes almost negligible, when I use an L3 cache and >> turn on the prefetchers at the L2. In case of 3 level cache hierarchy, this >> equation holds true: >> pf_issued at the L2s = Pf hits at the L3 + Pf requests >> responded by the DRAM >> >> Here are the actual numbers for a 4 core setup with PARSEC benchmarks: >> >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 130935 # Number of read requests responded to by this >> memory >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 148226 # number of hwpf issued >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 346384 # Number of read requests responded to by this >> memory >> canneal_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 354459 # number of hwpf issued >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 2937725 # Number of read requests responded to by >> this memory >> dedup_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 3251216 # number of hwpf issued >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 3443990 # Number of read requests responded to by >> this memory >> facesim_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 3639481 # number of hwpf issued >> ferret_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 2141141 # Number of read requests responded to by >> this memory >> ferret_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 12161061 # number of hwpf issued >> freqmine_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 115625 # Number of read requests responded to by this >> memory >> freqmine_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 168290 # number of hwpf issued >> rtview_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 590098 # Number of read requests responded to by this >> memory >> rtview_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 863745 # number of hwpf issued >> swaptions_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 20308 # Number of read requests responded to by this >> memory >> swaptions_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 16689699 # number of hwpf issued >> vips_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 3361213 # Number of read requests responded to by >> this memory >> vips_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 3542000 # number of hwpf issued >> x264_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >> 1060135 # Number of read requests responded to by >> this memory >> x264_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >> 1805457 # number of hwpf issued >> >> >> >> >> >> >> >> -- >> >> >> *thanks&regards * >> *BISWABANDAN* >> http://www.cse.iitm.ac.in/~biswa/ >> >> “We might fall down, but we will never lay down. We might not be the >> best, but we will beat the best! We might not be at the top, but we will >> rise.” >> >> >> >> -- IMPORTANT NOTICE: The contents of this email and any attachments are >> confidential and may also be privileged. If you are not the intended >> recipient, please notify the sender immediately and do not disclose the >> contents to any other person, use it for any purpose, or store or copy the >> information in any medium. Thank you. >> >> ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, >> Registered in England & Wales, Company No: 2557590 >> ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, >> Registered in England & Wales, Company No: 2548782 >> > > > > -- > > > *thanks&regards* > *BISWABANDAN* > http://www.cse.iitm.ac.in/~biswa/ > > “We might fall down, but we will never lay down. We might not be the best, > but we will beat the best! We might not be at the top, but we will rise.” > > > > _______________________________________________ > gem5-users mailing list > gem5-users(a)gem5.org > http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users >

Hi All,
         For a 2 level cache hierarchy with DDR3 (11-11-11) as the

Thanks Amin. I am using the latest gem5. Before generating the prefetch requests, the prefetcher does check whether the corresponding blk address (of the to be prefetched address) is already in the MSHR or in the MissQueue. So I don't think it is about the multiple packets. On Mon, Jun 30, 2014 at 9:41 PM, Amin Farmahini <aminfar(a)gmail.com> wrote: > Just something to note that the prefetcher in previous versions of gem5 > does not operate as it is supposed. A couple of patches were pushed in > early this year to fix this issue, but I never checked to see if they > really fixed the issue. The issue, if I remember right, is that prefetcher > may generate multiple requests per cache line which is basically merged to > one request in the DRAM read queue. > I am not sure this directly relates to your problem... > > Thanks, > Amin > > > On Mon, Jun 30, 2014 at 7:50 AM, biswabandan panda via gem5-users < > gem5-users(a)gem5.org> wrote: > >> Hi Andreas, >> Thanks for the prompt response. Actually DRAM readQ >> is never filled up completely. >> >> Here are some of the stats: >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry >> 0 # Number of times read queue was full causing retry >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 >> 29760 # What read queue length does an incoming req >> see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 >> 23193 # What read queue length does an incoming req >> see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 >> 18571 # What read queue length does an incoming req >> see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 >> 15509 # What read queue length does an incoming req >> see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 >> 12970 # What read queue length does an incoming req >> see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 >> 11623 # What read queue length does an incoming req >> see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 >> 10339 # What read queue length does an incoming req >> see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 >> 9681 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 >> 3538 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 >> 2324 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 >> 1782 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 >> 1758 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 >> 1112 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 >> 957 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 >> 822 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 >> 710 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 >> 0 # What read queue length does an incoming req see >> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry >> 0 # Number of times read queue was full causing retry >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 >> 1820343 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 >> 820262 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 >> 214278 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 >> 59606 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 >> 47136 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 >> 46901 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 >> 45035 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 >> 43390 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 >> 38898 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 >> 22687 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 >> 11365 # What read queue length does an incoming req >> see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 >> 6793 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 >> 3105 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 >> 1925 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 >> 1095 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 >> 654 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 >> 0 # What read queue length does an incoming req see >> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry >> 0 # Number of times read queue was full causing retry >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 >> 601147 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 >> 466636 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 >> 397007 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 >> 320737 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 >> 273422 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 >> 235104 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 >> 215016 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 >> 204993 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 >> 160861 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 >> 104896 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 >> 78910 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 >> 84566 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 >> 42656 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 >> 32575 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 >> 24300 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 >> 18088 # What read queue length does an incoming req >> see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 >> 0 # What read queue length does an incoming req see >> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.numRdRetry >> 0 # Number of times read queue was full causing retry >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::0 >> 923820 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::1 >> 570722 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::2 >> 412947 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::3 >> 336858 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::4 >> 278383 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::5 >> 242649 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::6 >> 218409 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::7 >> 205175 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::8 >> 176151 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::9 >> 159665 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::10 >> 137869 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::11 >> 129144 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::12 >> 75569 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::13 >> 52159 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::14 >> 35059 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::15 >> 24994 # What read queue length does an incoming req >> see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::16 >> 9 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::17 >> 6 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::18 >> 1 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::19 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::20 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::21 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::22 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::23 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::24 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::25 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::26 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::27 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::28 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::29 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::30 >> 0 # What read queue length does an incoming req see >> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.rdQLenPdf::31 >> >> >> On Mon, Jun 30, 2014 at 6:29 PM, Andreas Hansson <Andreas.Hansson(a)arm.com >> > wrote: >> >>> Hi Biswabandan, >>> >>> Is there any chance you could dig a bit deeper here? >>> >>> My suspicion is that the DRAM read queue is filling up and forcing >>> packets to wait. The controller only counts reads when they are actually >>> serviced. Perhaps the prefetcher counts issues prefetches as it creates the >>> packets, even if they are not successfully sent? Check numRdRetry and the >>> rdQLenPdf. >>> >>> Thanks, >>> >>> Andreas >>> >>> From: biswabandan panda via gem5-users <gem5-users(a)gem5.org> >>> Reply-To: biswabandan panda <biswa.uce(a)gmail.com>, gem5 users mailing >>> list <gem5-users(a)gem5.org> >>> Date: Monday, 30 June 2014 12:19 >>> To: gem5 users mailing list <gem5-users(a)gem5.org> >>> Subject: [gem5-users] DRAM Controller and Prefetchers >>> >>> Hi All, >>> For a 2 level cache hierarchy with DDR3 (11-11-11) as the >>> DRAM specs, the number of prefetch requests issued by the L2 prefetcher is >>> high (very high in some cases) as compared to the number of prefetch >>> requests responded by the DRAM controller. Any thoughts on it. >>> >>> The difference becomes almost negligible, when I use an L3 cache and >>> turn on the prefetchers at the L2. In case of 3 level cache hierarchy, this >>> equation holds true: >>> pf_issued at the L2s = Pf hits at the L3 + Pf requests >>> responded by the DRAM >>> >>> Here are the actual numbers for a 4 core setup with PARSEC benchmarks: >>> >>> blackscholes_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 130935 # Number of read requests responded to by this >>> memory >>> blackscholes_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 148226 # number of hwpf issued >>> canneal_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 346384 # Number of read requests responded to by this >>> memory >>> canneal_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 354459 # number of hwpf issued >>> dedup_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 2937725 # Number of read requests responded to by >>> this memory >>> dedup_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 3251216 # number of hwpf issued >>> facesim_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 3443990 # Number of read requests responded to by >>> this memory >>> facesim_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 3639481 # number of hwpf issued >>> ferret_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 2141141 # Number of read requests responded to by >>> this memory >>> ferret_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 12161061 # number of hwpf issued >>> freqmine_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 115625 # Number of read requests responded to by this >>> memory >>> freqmine_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 168290 # number of hwpf issued >>> rtview_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 590098 # Number of read requests responded to by this >>> memory >>> rtview_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 863745 # number of hwpf issued >>> swaptions_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 20308 # Number of read requests responded to by this >>> memory >>> swaptions_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 16689699 # number of hwpf issued >>> vips_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 3361213 # Number of read requests responded to by >>> this memory >>> vips_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 3542000 # number of hwpf issued >>> x264_4c_simlarge_ckpts.rcS.txt:system.mem_ctrls.num_reads::l2.prefetcher >>> 1060135 # Number of read requests responded to by >>> this memory >>> x264_4c_simlarge_ckpts.rcS.txt:system.l2.prefetcher.prefetcher.num_hwpf_issued >>> 1805457 # number of hwpf issued >>> >>> >>> >>> >>> >>> >>> >>> -- >>> >>> >>> *thanks&regards * >>> *BISWABANDAN* >>> http://www.cse.iitm.ac.in/~biswa/ >>> >>> “We might fall down, but we will never lay down. We might not be the >>> best, but we will beat the best! We might not be at the top, but we will >>> rise.” >>> >>> >>> >>> -- IMPORTANT NOTICE: The contents of this email and any attachments are >>> confidential and may also be privileged. If you are not the intended >>> recipient, please notify the sender immediately and do not disclose the >>> contents to any other person, use it for any purpose, or store or copy the >>> information in any medium. Thank you. >>> >>> ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, >>> Registered in England & Wales, Company No: 2557590 >>> ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 >>> 9NJ, Registered in England & Wales, Company No: 2548782 >>> >> >> >> >> -- >> >> >> *thanks&regards* >> *BISWABANDAN* >> http://www.cse.iitm.ac.in/~biswa/ >> >> “We might fall down, but we will never lay down. We might not be the >> best, but we will beat the best! We might not be at the top, but we will >> rise.” >> >> >> >> _______________________________________________ >> gem5-users mailing list >> gem5-users(a)gem5.org >> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users >> > > -- *thanks&regards* *BISWABANDAN* http://www.cse.iitm.ac.in/~biswa/ “We might fall down, but we will never lay down. We might not be the best, but we will beat the best! We might not be at the top, but we will rise.”