Dear gem5 community, I have successfully cross-compile the HPCG benchmark for RISCV (Serial version, without MPI and OpenMP). While it working properly in gem5 SE mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 --nz=16 --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv image and put it). Can you help me please? In addition, I used the RISCV Ubuntu image (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), I installed the gcc compiler, compile it (through qemu) and I get wrong results too. Here is the Makefile which I use, the hpcg executable for RISCV (xhpcg), and a video that shows the results (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). P.S. I use the latest gem5 version. Thank you in advance! :) Best regards, Nikos

I'm going to need a bit more information to help: 1. In what way have you modified ./configs/example/gem5_library/riscv-fs.py? Can you attach the script here? 2. What error are you getting or in what way are the results invalid? - Dr. Bobby R. Bruce Room 3050, Kemper Hall, UC Davis Davis, CA, 95616 web: https://www.bobbybruce.net On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < ntampouratzis@ece.auth.gr> wrote: > > Dear gem5 community, > > I have successfully cross-compile the HPCG benchmark for RISCV (Serial > version, without MPI and OpenMP). While it working properly in gem5 SE > mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results > ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 --nz=16 > --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS > simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results > ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv image > and put it). > > Can you help me please? > > In addition, I used the RISCV Ubuntu image > (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), > I installed the gcc compiler, compile it (through qemu) and I get > wrong results too. > > Here is the Makefile which I use, the hpcg executable for RISCV > (xhpcg), and a video that shows the results > (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). > > P.S. I use the latest gem5 version. > > Thank you in advance! :) > > Best regards, > Nikos > _______________________________________________ > gem5-users mailing list -- gem5-users@gem5.org > To unsubscribe send an email to gem5-users-leave@gem5.org >

Dear Bobby, 1) I use the original riscv-fs.py which is provided in the latest gem5 release. I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results ./configs/example/gem5_library/riscv-fs.py) in order to download the riscv-bootloader-vmlinux-5.10 and riscv-disk-img. After this I mount the riscv-disk-img (sudo mount -o loop riscv-disk-img /mnt), put the xhpcg executable and I do the following changes in riscv-fs.py to boot the riscv-disk-img with executable: image = CustomDiskImageResource( local_path = "/home/cossim/.cache/gem5/riscv-disk-img", ) # Set the Full System workload. board.set_kernel_disk_workload( kernel=Resource("riscv-bootloader-vmlinux-5.10"), disk_image=image, ) Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py I change the last line to "return ["console=ttyS0", "root={root_value}", "rw"]" in order to allow the write permissions in the image. 2) The HPCG benchmark after some iterations calculates if the results are valid or not valid. In the case of FS it gives invalid results. As I see from the results, one (at least) problem is that produces different results in each HPCG execution (with the same configuration). Here is the HPCG output and riscv-fs.py (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the results in the video if you use the xhpcg executable (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) Please help me in order to solve it! Finally, I get invalid results in the HPL benchmark in FS mode too. Best regards, Nikos Quoting Bobby Bruce <bbruce@ucdavis.edu>: > I'm going to need a bit more information to help: > > 1. In what way have you modified > ./configs/example/gem5_library/riscv-fs.py? Can you attach the script here? > 2. What error are you getting or in what way are the results invalid? > > - > Dr. Bobby R. Bruce > Room 3050, > Kemper Hall, UC Davis > Davis, > CA, 95616 > > web: https://www.bobbybruce.net > > > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < > ntampouratzis@ece.auth.gr> wrote: > >> >> Dear gem5 community, >> >> I have successfully cross-compile the HPCG benchmark for RISCV (Serial >> version, without MPI and OpenMP). While it working properly in gem5 SE >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 --nz=16 >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv image >> and put it). >> >> Can you help me please? >> >> In addition, I used the RISCV Ubuntu image >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), >> I installed the gcc compiler, compile it (through qemu) and I get >> wrong results too. >> >> Here is the Makefile which I use, the hpcg executable for RISCV >> (xhpcg), and a video that shows the results >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). >> >> P.S. I use the latest gem5 version. >> >> Thank you in advance! :) >> >> Best regards, >> Nikos >> _______________________________________________ >> gem5-users mailing list -- gem5-users@gem5.org >> To unsubscribe send an email to gem5-users-leave@gem5.org >>

Hi Nikos, I notice you said the following in your original email: In addition, I used the RISCV Ubuntu image > (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), > I installed the gcc compiler, compile it (through qemu) and I get > wrong results too. Is this saying you get the wrong results is QEMU? If so, the bug is in GCC or the HPCG workload, not in gem5. If not, I would test in QEMU to make sure the binary works there. Another way you could test to see if the problem is your binary or gem5 would be to run it on real hardware. We have access to some RISC-V hardware here at UC Davis, if you don't have access to it. Cheers, Jason On Tue, Sep 20, 2022 at 12:58 AM Νικόλαος Ταμπουρατζής < ntampouratzis@ece.auth.gr> wrote: > Dear Bobby, > > 1) I use the original riscv-fs.py which is provided in the latest gem5 > release. > I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results > ./configs/example/gem5_library/riscv-fs.py) in order to download the > riscv-bootloader-vmlinux-5.10 and riscv-disk-img. > After this I mount the riscv-disk-img (sudo mount -o loop > riscv-disk-img /mnt), put the xhpcg executable and I do the following > changes in riscv-fs.py to boot the riscv-disk-img with executable: > > image = CustomDiskImageResource( > local_path = "/home/cossim/.cache/gem5/riscv-disk-img", > ) > > # Set the Full System workload. > board.set_kernel_disk_workload( > kernel=Resource("riscv-bootloader-vmlinux-5.10"), > disk_image=image, > ) > > Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py > I change the last line to "return ["console=ttyS0", > "root={root_value}", "rw"]" in order to allow the write permissions in > the image. > > > 2) The HPCG benchmark after some iterations calculates if the results > are valid or not valid. In the case of FS it gives invalid results. As > I see from the results, one (at least) problem is that produces > different results in each HPCG execution (with the same configuration). > > Here is the HPCG output and riscv-fs.py > (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the > results in the video if you use the xhpcg executable > (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) > > Please help me in order to solve it! > > Finally, I get invalid results in the HPL benchmark in FS mode too. > > Best regards, > Nikos > > > Quoting Bobby Bruce <bbruce@ucdavis.edu>: > > > I'm going to need a bit more information to help: > > > > 1. In what way have you modified > > ./configs/example/gem5_library/riscv-fs.py? Can you attach the script > here? > > 2. What error are you getting or in what way are the results invalid? > > > > - > > Dr. Bobby R. Bruce > > Room 3050, > > Kemper Hall, UC Davis > > Davis, > > CA, 95616 > > > > web: https://www.bobbybruce.net > > > > > > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < > > ntampouratzis@ece.auth.gr> wrote: > > > >> > >> Dear gem5 community, > >> > >> I have successfully cross-compile the HPCG benchmark for RISCV (Serial > >> version, without MPI and OpenMP). While it working properly in gem5 SE > >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results > >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 --nz=16 > >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS > >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results > >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv image > >> and put it). > >> > >> Can you help me please? > >> > >> In addition, I used the RISCV Ubuntu image > >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), > >> I installed the gcc compiler, compile it (through qemu) and I get > >> wrong results too. > >> > >> Here is the Makefile which I use, the hpcg executable for RISCV > >> (xhpcg), and a video that shows the results > >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). > >> > >> P.S. I use the latest gem5 version. > >> > >> Thank you in advance! :) > >> > >> Best regards, > >> Nikos > >> _______________________________________________ > >> gem5-users mailing list -- gem5-users@gem5.org > >> To unsubscribe send an email to gem5-users-leave@gem5.org > >> > > > _______________________________________________ > gem5-users mailing list -- gem5-users@gem5.org > To unsubscribe send an email to gem5-users-leave@gem5.org >

Dear Bobby, In QEMU I get the same (correct) results that I get in SE mode simulation. I get invalid results in FS simulation (in both riscv-fs.py and riscv-ubuntu-run.py). I cannot access real RISCV hardware at this moment, however, if you want you may execute my xhpcg binary (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) with the following configuration: ./xhpcg --nx=16 --ny=16 --nz=16 --npx=1 --npy=1 --npz=1 --rt=0.1 Please let me know if you have any updates! Best regards, Nikos Quoting Jason Lowe-Power <jason@lowepower.com>: > Hi Nikos, > > I notice you said the following in your original email: > > In addition, I used the RISCV Ubuntu image >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), >> I installed the gcc compiler, compile it (through qemu) and I get >> wrong results too. > > > Is this saying you get the wrong results is QEMU? If so, the bug is in GCC > or the HPCG workload, not in gem5. If not, I would test in QEMU to make > sure the binary works there. Another way you could test to see if the > problem is your binary or gem5 would be to run it on real hardware. We have > access to some RISC-V hardware here at UC Davis, if you don't have access > to it. > > Cheers, > Jason > > On Tue, Sep 20, 2022 at 12:58 AM Νικόλαος Ταμπουρατζής < > ntampouratzis@ece.auth.gr> wrote: > >> Dear Bobby, >> >> 1) I use the original riscv-fs.py which is provided in the latest gem5 >> release. >> I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results >> ./configs/example/gem5_library/riscv-fs.py) in order to download the >> riscv-bootloader-vmlinux-5.10 and riscv-disk-img. >> After this I mount the riscv-disk-img (sudo mount -o loop >> riscv-disk-img /mnt), put the xhpcg executable and I do the following >> changes in riscv-fs.py to boot the riscv-disk-img with executable: >> >> image = CustomDiskImageResource( >> local_path = "/home/cossim/.cache/gem5/riscv-disk-img", >> ) >> >> # Set the Full System workload. >> board.set_kernel_disk_workload( >> kernel=Resource("riscv-bootloader-vmlinux-5.10"), >> disk_image=image, >> ) >> >> Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py >> I change the last line to "return ["console=ttyS0", >> "root={root_value}", "rw"]" in order to allow the write permissions in >> the image. >> >> >> 2) The HPCG benchmark after some iterations calculates if the results >> are valid or not valid. In the case of FS it gives invalid results. As >> I see from the results, one (at least) problem is that produces >> different results in each HPCG execution (with the same configuration). >> >> Here is the HPCG output and riscv-fs.py >> (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the >> results in the video if you use the xhpcg executable >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) >> >> Please help me in order to solve it! >> >> Finally, I get invalid results in the HPL benchmark in FS mode too. >> >> Best regards, >> Nikos >> >> >> Quoting Bobby Bruce <bbruce@ucdavis.edu>: >> >> > I'm going to need a bit more information to help: >> > >> > 1. In what way have you modified >> > ./configs/example/gem5_library/riscv-fs.py? Can you attach the script >> here? >> > 2. What error are you getting or in what way are the results invalid? >> > >> > - >> > Dr. Bobby R. Bruce >> > Room 3050, >> > Kemper Hall, UC Davis >> > Davis, >> > CA, 95616 >> > >> > web: https://www.bobbybruce.net >> > >> > >> > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < >> > ntampouratzis@ece.auth.gr> wrote: >> > >> >> >> >> Dear gem5 community, >> >> >> >> I have successfully cross-compile the HPCG benchmark for RISCV (Serial >> >> version, without MPI and OpenMP). While it working properly in gem5 SE >> >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results >> >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 --nz=16 >> >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS >> >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results >> >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv image >> >> and put it). >> >> >> >> Can you help me please? >> >> >> >> In addition, I used the RISCV Ubuntu image >> >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), >> >> I installed the gcc compiler, compile it (through qemu) and I get >> >> wrong results too. >> >> >> >> Here is the Makefile which I use, the hpcg executable for RISCV >> >> (xhpcg), and a video that shows the results >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). >> >> >> >> P.S. I use the latest gem5 version. >> >> >> >> Thank you in advance! :) >> >> >> >> Best regards, >> >> Nikos >> >> _______________________________________________ >> >> gem5-users mailing list -- gem5-users@gem5.org >> >> To unsubscribe send an email to gem5-users-leave@gem5.org >> >> >> >> >> _______________________________________________ >> gem5-users mailing list -- gem5-users@gem5.org >> To unsubscribe send an email to gem5-users-leave@gem5.org >>

That's quite odd that it works in SE mode but not FS mode! I would suggest running with --debug-flags=Exec for both and then perform a diff to see how they differ. Cheers, Jason On Tue, Sep 20, 2022 at 2:45 PM Νικόλαος Ταμπουρατζής < ntampouratzis@ece.auth.gr> wrote: > Dear Bobby, > > In QEMU I get the same (correct) results that I get in SE mode > simulation. I get invalid results in FS simulation (in both > riscv-fs.py and riscv-ubuntu-run.py). I cannot access real RISCV > hardware at this moment, however, if you want you may execute my xhpcg > binary (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) with the > following configuration: > > ./xhpcg --nx=16 --ny=16 --nz=16 --npx=1 --npy=1 --npz=1 --rt=0.1 > > Please let me know if you have any updates! > > Best regards, > Nikos > > > Quoting Jason Lowe-Power <jason@lowepower.com>: > > > Hi Nikos, > > > > I notice you said the following in your original email: > > > > In addition, I used the RISCV Ubuntu image > >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), > >> I installed the gcc compiler, compile it (through qemu) and I get > >> wrong results too. > > > > > > Is this saying you get the wrong results is QEMU? If so, the bug is in > GCC > > or the HPCG workload, not in gem5. If not, I would test in QEMU to make > > sure the binary works there. Another way you could test to see if the > > problem is your binary or gem5 would be to run it on real hardware. We > have > > access to some RISC-V hardware here at UC Davis, if you don't have access > > to it. > > > > Cheers, > > Jason > > > > On Tue, Sep 20, 2022 at 12:58 AM Νικόλαος Ταμπουρατζής < > > ntampouratzis@ece.auth.gr> wrote: > > > >> Dear Bobby, > >> > >> 1) I use the original riscv-fs.py which is provided in the latest gem5 > >> release. > >> I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results > >> ./configs/example/gem5_library/riscv-fs.py) in order to download the > >> riscv-bootloader-vmlinux-5.10 and riscv-disk-img. > >> After this I mount the riscv-disk-img (sudo mount -o loop > >> riscv-disk-img /mnt), put the xhpcg executable and I do the following > >> changes in riscv-fs.py to boot the riscv-disk-img with executable: > >> > >> image = CustomDiskImageResource( > >> local_path = "/home/cossim/.cache/gem5/riscv-disk-img", > >> ) > >> > >> # Set the Full System workload. > >> board.set_kernel_disk_workload( > >> kernel=Resource("riscv-bootloader-vmlinux-5.10"), > >> disk_image=image, > >> ) > >> > >> Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py > >> I change the last line to "return ["console=ttyS0", > >> "root={root_value}", "rw"]" in order to allow the write permissions in > >> the image. > >> > >> > >> 2) The HPCG benchmark after some iterations calculates if the results > >> are valid or not valid. In the case of FS it gives invalid results. As > >> I see from the results, one (at least) problem is that produces > >> different results in each HPCG execution (with the same configuration). > >> > >> Here is the HPCG output and riscv-fs.py > >> (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the > >> results in the video if you use the xhpcg executable > >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) > >> > >> Please help me in order to solve it! > >> > >> Finally, I get invalid results in the HPL benchmark in FS mode too. > >> > >> Best regards, > >> Nikos > >> > >> > >> Quoting Bobby Bruce <bbruce@ucdavis.edu>: > >> > >> > I'm going to need a bit more information to help: > >> > > >> > 1. In what way have you modified > >> > ./configs/example/gem5_library/riscv-fs.py? Can you attach the script > >> here? > >> > 2. What error are you getting or in what way are the results invalid? > >> > > >> > - > >> > Dr. Bobby R. Bruce > >> > Room 3050, > >> > Kemper Hall, UC Davis > >> > Davis, > >> > CA, 95616 > >> > > >> > web: https://www.bobbybruce.net > >> > > >> > > >> > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < > >> > ntampouratzis@ece.auth.gr> wrote: > >> > > >> >> > >> >> Dear gem5 community, > >> >> > >> >> I have successfully cross-compile the HPCG benchmark for RISCV > (Serial > >> >> version, without MPI and OpenMP). While it working properly in gem5 > SE > >> >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results > >> >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 --nz=16 > >> >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS > >> >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results > >> >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv image > >> >> and put it). > >> >> > >> >> Can you help me please? > >> >> > >> >> In addition, I used the RISCV Ubuntu image > >> >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu > ), > >> >> I installed the gcc compiler, compile it (through qemu) and I get > >> >> wrong results too. > >> >> > >> >> Here is the Makefile which I use, the hpcg executable for RISCV > >> >> (xhpcg), and a video that shows the results > >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). > >> >> > >> >> P.S. I use the latest gem5 version. > >> >> > >> >> Thank you in advance! :) > >> >> > >> >> Best regards, > >> >> Nikos > >> >> _______________________________________________ > >> >> gem5-users mailing list -- gem5-users@gem5.org > >> >> To unsubscribe send an email to gem5-users-leave@gem5.org > >> >> > >> > >> > >> _______________________________________________ > >> gem5-users mailing list -- gem5-users@gem5.org > >> To unsubscribe send an email to gem5-users-leave@gem5.org > >> > > > _______________________________________________ > gem5-users mailing list -- gem5-users@gem5.org > To unsubscribe send an email to gem5-users-leave@gem5.org >

Dear Bobby, Iam trying to add --debug-flags=Exec (building the gem5 for gem5.opt not for gem5.fast which I had) but the debug traces exceed the 20GB (and it is not finished yet) for less than 1 simulated second. How can I reduce the size of the debug-flags (or set something more specific)? In contrast I build the HPCG benchmark with DHPCG_DEBUG flag. If you want, you can compare these two output files (hpcg20010909T014640_SE_Mode & HPCG-Benchmark_3.1_FS_Mode). As you can see, something goes wrong with the accuracy of calculations in FS mode (benchmark uses double precission). You can find the files here: http://kition.mhl.tuc.gr:8000/d/68d82f3533/ Best regards, Nikos Quoting Jason Lowe-Power <jason@lowepower.com>: > That's quite odd that it works in SE mode but not FS mode! > > I would suggest running with --debug-flags=Exec for both and then perform a > diff to see how they differ. > > Cheers, > Jason > > On Tue, Sep 20, 2022 at 2:45 PM Νικόλαος Ταμπουρατζής < > ntampouratzis@ece.auth.gr> wrote: > >> Dear Bobby, >> >> In QEMU I get the same (correct) results that I get in SE mode >> simulation. I get invalid results in FS simulation (in both >> riscv-fs.py and riscv-ubuntu-run.py). I cannot access real RISCV >> hardware at this moment, however, if you want you may execute my xhpcg >> binary (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) with the >> following configuration: >> >> ./xhpcg --nx=16 --ny=16 --nz=16 --npx=1 --npy=1 --npz=1 --rt=0.1 >> >> Please let me know if you have any updates! >> >> Best regards, >> Nikos >> >> >> Quoting Jason Lowe-Power <jason@lowepower.com>: >> >> > Hi Nikos, >> > >> > I notice you said the following in your original email: >> > >> > In addition, I used the RISCV Ubuntu image >> >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu), >> >> I installed the gcc compiler, compile it (through qemu) and I get >> >> wrong results too. >> > >> > >> > Is this saying you get the wrong results is QEMU? If so, the bug is in >> GCC >> > or the HPCG workload, not in gem5. If not, I would test in QEMU to make >> > sure the binary works there. Another way you could test to see if the >> > problem is your binary or gem5 would be to run it on real hardware. We >> have >> > access to some RISC-V hardware here at UC Davis, if you don't have access >> > to it. >> > >> > Cheers, >> > Jason >> > >> > On Tue, Sep 20, 2022 at 12:58 AM Νικόλαος Ταμπουρατζής < >> > ntampouratzis@ece.auth.gr> wrote: >> > >> >> Dear Bobby, >> >> >> >> 1) I use the original riscv-fs.py which is provided in the latest gem5 >> >> release. >> >> I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results >> >> ./configs/example/gem5_library/riscv-fs.py) in order to download the >> >> riscv-bootloader-vmlinux-5.10 and riscv-disk-img. >> >> After this I mount the riscv-disk-img (sudo mount -o loop >> >> riscv-disk-img /mnt), put the xhpcg executable and I do the following >> >> changes in riscv-fs.py to boot the riscv-disk-img with executable: >> >> >> >> image = CustomDiskImageResource( >> >> local_path = "/home/cossim/.cache/gem5/riscv-disk-img", >> >> ) >> >> >> >> # Set the Full System workload. >> >> board.set_kernel_disk_workload( >> >> kernel=Resource("riscv-bootloader-vmlinux-5.10"), >> >> disk_image=image, >> >> ) >> >> >> >> Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py >> >> I change the last line to "return ["console=ttyS0", >> >> "root={root_value}", "rw"]" in order to allow the write permissions in >> >> the image. >> >> >> >> >> >> 2) The HPCG benchmark after some iterations calculates if the results >> >> are valid or not valid. In the case of FS it gives invalid results. As >> >> I see from the results, one (at least) problem is that produces >> >> different results in each HPCG execution (with the same configuration). >> >> >> >> Here is the HPCG output and riscv-fs.py >> >> (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the >> >> results in the video if you use the xhpcg executable >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) >> >> >> >> Please help me in order to solve it! >> >> >> >> Finally, I get invalid results in the HPL benchmark in FS mode too. >> >> >> >> Best regards, >> >> Nikos >> >> >> >> >> >> Quoting Bobby Bruce <bbruce@ucdavis.edu>: >> >> >> >> > I'm going to need a bit more information to help: >> >> > >> >> > 1. In what way have you modified >> >> > ./configs/example/gem5_library/riscv-fs.py? Can you attach the script >> >> here? >> >> > 2. What error are you getting or in what way are the results invalid? >> >> > >> >> > - >> >> > Dr. Bobby R. Bruce >> >> > Room 3050, >> >> > Kemper Hall, UC Davis >> >> > Davis, >> >> > CA, 95616 >> >> > >> >> > web: https://www.bobbybruce.net >> >> > >> >> > >> >> > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < >> >> > ntampouratzis@ece.auth.gr> wrote: >> >> > >> >> >> >> >> >> Dear gem5 community, >> >> >> >> >> >> I have successfully cross-compile the HPCG benchmark for RISCV >> (Serial >> >> >> version, without MPI and OpenMP). While it working properly in gem5 >> SE >> >> >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results >> >> >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 --nz=16 >> >> >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS >> >> >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results >> >> >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv image >> >> >> and put it). >> >> >> >> >> >> Can you help me please? >> >> >> >> >> >> In addition, I used the RISCV Ubuntu image >> >> >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu >> ), >> >> >> I installed the gcc compiler, compile it (through qemu) and I get >> >> >> wrong results too. >> >> >> >> >> >> Here is the Makefile which I use, the hpcg executable for RISCV >> >> >> (xhpcg), and a video that shows the results >> >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). >> >> >> >> >> >> P.S. I use the latest gem5 version. >> >> >> >> >> >> Thank you in advance! :) >> >> >> >> >> >> Best regards, >> >> >> Nikos >> >> >> _______________________________________________ >> >> >> gem5-users mailing list -- gem5-users@gem5.org >> >> >> To unsubscribe send an email to gem5-users-leave@gem5.org >> >> >> >> >> >> >> >> >> _______________________________________________ >> >> gem5-users mailing list -- gem5-users@gem5.org >> >> To unsubscribe send an email to gem5-users-leave@gem5.org >> >> >> >> >> _______________________________________________ >> gem5-users mailing list -- gem5-users@gem5.org >> To unsubscribe send an email to gem5-users-leave@gem5.org >>

Hi Nikos, You can use --debug-start to start the debugging after some number of ticks. Also, I would expect that the difference should come up quickly, so no need to run the program to the end. For the FS mode one, you will want to just start the trace as the application starts. This could be a bit of a pain. I'm not really sure what fundamentally could be different. FS and SE mode use the exact same code for executing instructions, so I don't think that's the problem. Have you tried running for smaller inputs or just one iteration? Jason On Wed, Sep 21, 2022 at 9:04 AM Νικόλαος Ταμπουρατζής < ntampouratzis@ece.auth.gr> wrote: > Dear Bobby, > > Iam trying to add --debug-flags=Exec (building the gem5 for gem5.opt > not for gem5.fast which I had) but the debug traces exceed the 20GB > (and it is not finished yet) for less than 1 simulated second. How can > I reduce the size of the debug-flags (or set something more specific)? > > In contrast I build the HPCG benchmark with DHPCG_DEBUG flag. If you > want, you can compare these two output files > (hpcg20010909T014640_SE_Mode & HPCG-Benchmark_3.1_FS_Mode). As you can > see, something goes wrong with the accuracy of calculations in FS mode > (benchmark uses double precission). You can find the files here: > http://kition.mhl.tuc.gr:8000/d/68d82f3533/ > > Best regards, > Nikos > > Quoting Jason Lowe-Power <jason@lowepower.com>: > > > That's quite odd that it works in SE mode but not FS mode! > > > > I would suggest running with --debug-flags=Exec for both and then > perform a > > diff to see how they differ. > > > > Cheers, > > Jason > > > > On Tue, Sep 20, 2022 at 2:45 PM Νικόλαος Ταμπουρατζής < > > ntampouratzis@ece.auth.gr> wrote: > > > >> Dear Bobby, > >> > >> In QEMU I get the same (correct) results that I get in SE mode > >> simulation. I get invalid results in FS simulation (in both > >> riscv-fs.py and riscv-ubuntu-run.py). I cannot access real RISCV > >> hardware at this moment, however, if you want you may execute my xhpcg > >> binary (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) with the > >> following configuration: > >> > >> ./xhpcg --nx=16 --ny=16 --nz=16 --npx=1 --npy=1 --npz=1 --rt=0.1 > >> > >> Please let me know if you have any updates! > >> > >> Best regards, > >> Nikos > >> > >> > >> Quoting Jason Lowe-Power <jason@lowepower.com>: > >> > >> > Hi Nikos, > >> > > >> > I notice you said the following in your original email: > >> > > >> > In addition, I used the RISCV Ubuntu image > >> >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu > ), > >> >> I installed the gcc compiler, compile it (through qemu) and I get > >> >> wrong results too. > >> > > >> > > >> > Is this saying you get the wrong results is QEMU? If so, the bug is in > >> GCC > >> > or the HPCG workload, not in gem5. If not, I would test in QEMU to > make > >> > sure the binary works there. Another way you could test to see if the > >> > problem is your binary or gem5 would be to run it on real hardware. We > >> have > >> > access to some RISC-V hardware here at UC Davis, if you don't have > access > >> > to it. > >> > > >> > Cheers, > >> > Jason > >> > > >> > On Tue, Sep 20, 2022 at 12:58 AM Νικόλαος Ταμπουρατζής < > >> > ntampouratzis@ece.auth.gr> wrote: > >> > > >> >> Dear Bobby, > >> >> > >> >> 1) I use the original riscv-fs.py which is provided in the latest > gem5 > >> >> release. > >> >> I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results > >> >> ./configs/example/gem5_library/riscv-fs.py) in order to download the > >> >> riscv-bootloader-vmlinux-5.10 and riscv-disk-img. > >> >> After this I mount the riscv-disk-img (sudo mount -o loop > >> >> riscv-disk-img /mnt), put the xhpcg executable and I do the following > >> >> changes in riscv-fs.py to boot the riscv-disk-img with executable: > >> >> > >> >> image = CustomDiskImageResource( > >> >> local_path = "/home/cossim/.cache/gem5/riscv-disk-img", > >> >> ) > >> >> > >> >> # Set the Full System workload. > >> >> board.set_kernel_disk_workload( > >> >> kernel=Resource("riscv-bootloader-vmlinux-5.10"), > >> >> disk_image=image, > >> >> ) > >> >> > >> >> Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py > >> >> I change the last line to "return ["console=ttyS0", > >> >> "root={root_value}", "rw"]" in order to allow the write permissions > in > >> >> the image. > >> >> > >> >> > >> >> 2) The HPCG benchmark after some iterations calculates if the results > >> >> are valid or not valid. In the case of FS it gives invalid results. > As > >> >> I see from the results, one (at least) problem is that produces > >> >> different results in each HPCG execution (with the same > configuration). > >> >> > >> >> Here is the HPCG output and riscv-fs.py > >> >> (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the > >> >> results in the video if you use the xhpcg executable > >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) > >> >> > >> >> Please help me in order to solve it! > >> >> > >> >> Finally, I get invalid results in the HPL benchmark in FS mode too. > >> >> > >> >> Best regards, > >> >> Nikos > >> >> > >> >> > >> >> Quoting Bobby Bruce <bbruce@ucdavis.edu>: > >> >> > >> >> > I'm going to need a bit more information to help: > >> >> > > >> >> > 1. In what way have you modified > >> >> > ./configs/example/gem5_library/riscv-fs.py? Can you attach the > script > >> >> here? > >> >> > 2. What error are you getting or in what way are the results > invalid? > >> >> > > >> >> > - > >> >> > Dr. Bobby R. Bruce > >> >> > Room 3050, > >> >> > Kemper Hall, UC Davis > >> >> > Davis, > >> >> > CA, 95616 > >> >> > > >> >> > web: https://www.bobbybruce.net > >> >> > > >> >> > > >> >> > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < > >> >> > ntampouratzis@ece.auth.gr> wrote: > >> >> > > >> >> >> > >> >> >> Dear gem5 community, > >> >> >> > >> >> >> I have successfully cross-compile the HPCG benchmark for RISCV > >> (Serial > >> >> >> version, without MPI and OpenMP). While it working properly in > gem5 > >> SE > >> >> >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results > >> >> >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 > --nz=16 > >> >> >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS > >> >> >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results > >> >> >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv > image > >> >> >> and put it). > >> >> >> > >> >> >> Can you help me please? > >> >> >> > >> >> >> In addition, I used the RISCV Ubuntu image > >> >> >> ( > https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu > >> ), > >> >> >> I installed the gcc compiler, compile it (through qemu) and I get > >> >> >> wrong results too. > >> >> >> > >> >> >> Here is the Makefile which I use, the hpcg executable for RISCV > >> >> >> (xhpcg), and a video that shows the results > >> >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). > >> >> >> > >> >> >> P.S. I use the latest gem5 version. > >> >> >> > >> >> >> Thank you in advance! :) > >> >> >> > >> >> >> Best regards, > >> >> >> Nikos > >> >> >> _______________________________________________ > >> >> >> gem5-users mailing list -- gem5-users@gem5.org > >> >> >> To unsubscribe send an email to gem5-users-leave@gem5.org > >> >> >> > >> >> > >> >> > >> >> _______________________________________________ > >> >> gem5-users mailing list -- gem5-users@gem5.org > >> >> To unsubscribe send an email to gem5-users-leave@gem5.org > >> >> > >> > >> > >> _______________________________________________ > >> gem5-users mailing list -- gem5-users@gem5.org > >> To unsubscribe send an email to gem5-users-leave@gem5.org > >> > > > _______________________________________________ > gem5-users mailing list -- gem5-users@gem5.org > To unsubscribe send an email to gem5-users-leave@gem5.org >

 int dim = 4096;

 double result;

 for (int iter = 0; iter < 2; iter++){
     result = 0;
     for (int i = 0; i < dim; i++){
         for (int j = 0; j < dim; j++){
             result += sqrt(i) * sqrt(j);
         }
     }
     printf("Result: %lf\n", result); //Result: 30530733453.127449
 }

Dear Jason, I am trying to use --debug-start but in FS mode it is very difficult to find the tick on which the application is started! However, I am writing the following very simple c++ program: #include <cmath> #include <stdio.h> int main(){ int dim = 4096; double result; for (int iter = 0; iter < 2; iter++){ result = 0; for (int i = 0; i < dim; i++){ for (int j = 0; j < dim; j++){ result += sqrt(i) * sqrt(j); } } printf("Result: %lf\n", result); //Result: 30530733453.127449 } } I cross-compile it using: riscv64-linux-gnu-g++ -static -O3 -o test_riscv test_riscv.cpp While in X86 (without cross-compilation of course), QEMU-RISCV, GEM5-SE the result is the same (30530733453.127449), in GEM5-FS the result is different! In addition, the result is also different between the 2 iterations. Please reproduce the error if you want in order to verify my result. Ηow can the issue be resolved? Thank you in advance! Best regards, Nikos Quoting Jason Lowe-Power <jason@lowepower.com>: > Hi Nikos, > > You can use --debug-start to start the debugging after some number of > ticks. Also, I would expect that the difference should come up quickly, so > no need to run the program to the end. > > For the FS mode one, you will want to just start the trace as the > application starts. This could be a bit of a pain. > > I'm not really sure what fundamentally could be different. FS and SE mode > use the exact same code for executing instructions, so I don't think that's > the problem. Have you tried running for smaller inputs or just one > iteration? > > Jason > > > > On Wed, Sep 21, 2022 at 9:04 AM Νικόλαος Ταμπουρατζής < > ntampouratzis@ece.auth.gr> wrote: > >> Dear Bobby, >> >> Iam trying to add --debug-flags=Exec (building the gem5 for gem5.opt >> not for gem5.fast which I had) but the debug traces exceed the 20GB >> (and it is not finished yet) for less than 1 simulated second. How can >> I reduce the size of the debug-flags (or set something more specific)? >> >> In contrast I build the HPCG benchmark with DHPCG_DEBUG flag. If you >> want, you can compare these two output files >> (hpcg20010909T014640_SE_Mode & HPCG-Benchmark_3.1_FS_Mode). As you can >> see, something goes wrong with the accuracy of calculations in FS mode >> (benchmark uses double precission). You can find the files here: >> http://kition.mhl.tuc.gr:8000/d/68d82f3533/ >> >> Best regards, >> Nikos >> >> Quoting Jason Lowe-Power <jason@lowepower.com>: >> >> > That's quite odd that it works in SE mode but not FS mode! >> > >> > I would suggest running with --debug-flags=Exec for both and then >> perform a >> > diff to see how they differ. >> > >> > Cheers, >> > Jason >> > >> > On Tue, Sep 20, 2022 at 2:45 PM Νικόλαος Ταμπουρατζής < >> > ntampouratzis@ece.auth.gr> wrote: >> > >> >> Dear Bobby, >> >> >> >> In QEMU I get the same (correct) results that I get in SE mode >> >> simulation. I get invalid results in FS simulation (in both >> >> riscv-fs.py and riscv-ubuntu-run.py). I cannot access real RISCV >> >> hardware at this moment, however, if you want you may execute my xhpcg >> >> binary (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) with the >> >> following configuration: >> >> >> >> ./xhpcg --nx=16 --ny=16 --nz=16 --npx=1 --npy=1 --npz=1 --rt=0.1 >> >> >> >> Please let me know if you have any updates! >> >> >> >> Best regards, >> >> Nikos >> >> >> >> >> >> Quoting Jason Lowe-Power <jason@lowepower.com>: >> >> >> >> > Hi Nikos, >> >> > >> >> > I notice you said the following in your original email: >> >> > >> >> > In addition, I used the RISCV Ubuntu image >> >> >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu >> ), >> >> >> I installed the gcc compiler, compile it (through qemu) and I get >> >> >> wrong results too. >> >> > >> >> > >> >> > Is this saying you get the wrong results is QEMU? If so, the bug is in >> >> GCC >> >> > or the HPCG workload, not in gem5. If not, I would test in QEMU to >> make >> >> > sure the binary works there. Another way you could test to see if the >> >> > problem is your binary or gem5 would be to run it on real hardware. We >> >> have >> >> > access to some RISC-V hardware here at UC Davis, if you don't have >> access >> >> > to it. >> >> > >> >> > Cheers, >> >> > Jason >> >> > >> >> > On Tue, Sep 20, 2022 at 12:58 AM Νικόλαος Ταμπουρατζής < >> >> > ntampouratzis@ece.auth.gr> wrote: >> >> > >> >> >> Dear Bobby, >> >> >> >> >> >> 1) I use the original riscv-fs.py which is provided in the latest >> gem5 >> >> >> release. >> >> >> I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results >> >> >> ./configs/example/gem5_library/riscv-fs.py) in order to download the >> >> >> riscv-bootloader-vmlinux-5.10 and riscv-disk-img. >> >> >> After this I mount the riscv-disk-img (sudo mount -o loop >> >> >> riscv-disk-img /mnt), put the xhpcg executable and I do the following >> >> >> changes in riscv-fs.py to boot the riscv-disk-img with executable: >> >> >> >> >> >> image = CustomDiskImageResource( >> >> >> local_path = "/home/cossim/.cache/gem5/riscv-disk-img", >> >> >> ) >> >> >> >> >> >> # Set the Full System workload. >> >> >> board.set_kernel_disk_workload( >> >> >> kernel=Resource("riscv-bootloader-vmlinux-5.10"), >> >> >> disk_image=image, >> >> >> ) >> >> >> >> >> >> Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py >> >> >> I change the last line to "return ["console=ttyS0", >> >> >> "root={root_value}", "rw"]" in order to allow the write permissions >> in >> >> >> the image. >> >> >> >> >> >> >> >> >> 2) The HPCG benchmark after some iterations calculates if the results >> >> >> are valid or not valid. In the case of FS it gives invalid results. >> As >> >> >> I see from the results, one (at least) problem is that produces >> >> >> different results in each HPCG execution (with the same >> configuration). >> >> >> >> >> >> Here is the HPCG output and riscv-fs.py >> >> >> (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the >> >> >> results in the video if you use the xhpcg executable >> >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) >> >> >> >> >> >> Please help me in order to solve it! >> >> >> >> >> >> Finally, I get invalid results in the HPL benchmark in FS mode too. >> >> >> >> >> >> Best regards, >> >> >> Nikos >> >> >> >> >> >> >> >> >> Quoting Bobby Bruce <bbruce@ucdavis.edu>: >> >> >> >> >> >> > I'm going to need a bit more information to help: >> >> >> > >> >> >> > 1. In what way have you modified >> >> >> > ./configs/example/gem5_library/riscv-fs.py? Can you attach the >> script >> >> >> here? >> >> >> > 2. What error are you getting or in what way are the results >> invalid? >> >> >> > >> >> >> > - >> >> >> > Dr. Bobby R. Bruce >> >> >> > Room 3050, >> >> >> > Kemper Hall, UC Davis >> >> >> > Davis, >> >> >> > CA, 95616 >> >> >> > >> >> >> > web: https://www.bobbybruce.net >> >> >> > >> >> >> > >> >> >> > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < >> >> >> > ntampouratzis@ece.auth.gr> wrote: >> >> >> > >> >> >> >> >> >> >> >> Dear gem5 community, >> >> >> >> >> >> >> >> I have successfully cross-compile the HPCG benchmark for RISCV >> >> (Serial >> >> >> >> version, without MPI and OpenMP). While it working properly in >> gem5 >> >> SE >> >> >> >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results >> >> >> >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 >> --nz=16 >> >> >> >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS >> >> >> >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results >> >> >> >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv >> image >> >> >> >> and put it). >> >> >> >> >> >> >> >> Can you help me please? >> >> >> >> >> >> >> >> In addition, I used the RISCV Ubuntu image >> >> >> >> ( >> https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu >> >> ), >> >> >> >> I installed the gcc compiler, compile it (through qemu) and I get >> >> >> >> wrong results too. >> >> >> >> >> >> >> >> Here is the Makefile which I use, the hpcg executable for RISCV >> >> >> >> (xhpcg), and a video that shows the results >> >> >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). >> >> >> >> >> >> >> >> P.S. I use the latest gem5 version. >> >> >> >> >> >> >> >> Thank you in advance! :) >> >> >> >> >> >> >> >> Best regards, >> >> >> >> Nikos >> >> >> >> _______________________________________________ >> >> >> >> gem5-users mailing list -- gem5-users@gem5.org >> >> >> >> To unsubscribe send an email to gem5-users-leave@gem5.org >> >> >> >> >> >> >> >> >> >> >> >> >> _______________________________________________ >> >> >> gem5-users mailing list -- gem5-users@gem5.org >> >> >> To unsubscribe send an email to gem5-users-leave@gem5.org >> >> >> >> >> >> >> >> >> _______________________________________________ >> >> gem5-users mailing list -- gem5-users@gem5.org >> >> To unsubscribe send an email to gem5-users-leave@gem5.org >> >> >> >> >> _______________________________________________ >> gem5-users mailing list -- gem5-users@gem5.org >> To unsubscribe send an email to gem5-users-leave@gem5.org >>

 int dim = 10;

 float result;

 for (int iter = 0; iter < 2; iter++){
     result = 0;
     for (int i = 0; i < dim; i++){
         for (int j = 0; j < dim; j++){
             float sq_i = sqrt(i);
             float sq_j = sqrt(j);
             result += sq_i * sq_j;
             printf("ITER: %d | i: %d | j: %d Result(i: %f | j: %f  

 int dim = 4096;

 double result;

 for (int iter = 0; iter < 2; iter++){
     result = 0;
     for (int i = 0; i < dim; i++){
         for (int j = 0; j < dim; j++){
             result += sqrt(i) * sqrt(j);
         }
     }
     printf("Result: %lf\n", result); //Result: 30530733453.127449
 }

Dear Jason, all, I am trying to find the accuracy problem with RISCV-FS and I observe that the problem is created (at least in my dummy example) because the variables (double) are set to zero in random simulated time (for this reason I get different results among executions of the same code). Specifically for the following dummy code: #include <cmath> #include <stdio.h> int main(){ int dim = 10; float result; for (int iter = 0; iter < 2; iter++){ result = 0; for (int i = 0; i < dim; i++){ for (int j = 0; j < dim; j++){ float sq_i = sqrt(i); float sq_j = sqrt(j); result += sq_i * sq_j; printf("ITER: %d | i: %d | j: %d Result(i: %f | j: %f | i*j: %f): %f\n", iter, i , j, sq_i, sq_j, sq_i * sq_j, result); } } printf("Final Result: %lf\n", result); } } The correct Final Result in both iterations is 372.721656. However, I get the following results in FS: ITER: 0 | i: 0 | j: 0 Result(i: 0.000000 | j: 0.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 1 Result(i: 0.000000 | j: 1.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 2 Result(i: 0.000000 | j: 1.414214 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 3 Result(i: 0.000000 | j: 1.732051 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 4 Result(i: 0.000000 | j: 2.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 5 Result(i: 0.000000 | j: 2.236068 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 6 Result(i: 0.000000 | j: 2.449490 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 7 Result(i: 0.000000 | j: 2.645751 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 8 Result(i: 0.000000 | j: 2.828427 | i*j: 0.000000): 0.000000 ITER: 0 | i: 0 | j: 9 Result(i: 0.000000 | j: 3.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 0 Result(i: 1.000000 | j: 0.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 1 Result(i: 1.000000 | j: 1.000000 | i*j: 1.000000): 1.000000 ITER: 0 | i: 1 | j: 2 Result(i: 1.000000 | j: 1.414214 | i*j: 1.414214): 2.414214 ITER: 0 | i: 1 | j: 3 Result(i: 1.000000 | j: 1.732051 | i*j: 1.732051): 4.146264 ITER: 0 | i: 1 | j: 4 Result(i: 0.000000 | j: 2.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 5 Result(i: 0.000000 | j: 2.236068 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 6 Result(i: 0.000000 | j: 2.449490 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 7 Result(i: 0.000000 | j: 2.645751 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 8 Result(i: 0.000000 | j: 2.828427 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 9 Result(i: 0.000000 | j: 3.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 2 | j: 0 Result(i: 1.414214 | j: 0.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 2 | j: 1 Result(i: 1.414214 | j: 1.000000 | i*j: 1.414214): 1.414214 ITER: 0 | i: 2 | j: 2 Result(i: 1.414214 | j: 1.414214 | i*j: 2.000000): 3.414214 ITER: 0 | i: 2 | j: 3 Result(i: 1.414214 | j: 1.732051 | i*j: 2.449490): 5.863703 ITER: 0 | i: 2 | j: 4 Result(i: 1.414214 | j: 2.000000 | i*j: 2.828427): 8.692130 ITER: 0 | i: 2 | j: 5 Result(i: 1.414214 | j: 2.236068 | i*j: 3.162278): 11.854408 ITER: 0 | i: 2 | j: 6 Result(i: 1.414214 | j: 2.449490 | i*j: 3.464102): 15.318510 ITER: 0 | i: 2 | j: 7 Result(i: 1.414214 | j: 2.645751 | i*j: 3.741657): 19.060167 ITER: 0 | i: 2 | j: 8 Result(i: 1.414214 | j: 2.828427 | i*j: 4.000000): 23.060167 ITER: 0 | i: 2 | j: 9 Result(i: 1.414214 | j: 3.000000 | i*j: 4.242641): 27.302808 ITER: 0 | i: 3 | j: 0 Result(i: 1.732051 | j: 0.000000 | i*j: 0.000000): 27.302808 ITER: 0 | i: 3 | j: 1 Result(i: 1.732051 | j: 1.000000 | i*j: 1.732051): 29.034859 ITER: 0 | i: 3 | j: 2 Result(i: 1.732051 | j: 1.414214 | i*j: 2.449490): 31.484348 ITER: 0 | i: 3 | j: 3 Result(i: 1.732051 | j: 1.732051 | i*j: 3.000000): 34.484348 ITER: 0 | i: 3 | j: 4 Result(i: 1.732051 | j: 2.000000 | i*j: 3.464102): 37.948450 ITER: 0 | i: 3 | j: 5 Result(i: 1.732051 | j: 2.236068 | i*j: 3.872983): 41.821433 ITER: 0 | i: 3 | j: 6 Result(i: 1.732051 | j: 2.449490 | i*j: 4.242641): 46.064074 ITER: 0 | i: 3 | j: 7 Result(i: 1.732051 | j: 2.645751 | i*j: 4.582576): 50.646650 ITER: 0 | i: 3 | j: 8 Result(i: 1.732051 | j: 2.828427 | i*j: 4.898979): 55.545629 ITER: 0 | i: 3 | j: 9 Result(i: 1.732051 | j: 3.000000 | i*j: 5.196152): 60.741782 ITER: 0 | i: 4 | j: 0 Result(i: 2.000000 | j: 0.000000 | i*j: 0.000000): 60.741782 ITER: 0 | i: 4 | j: 1 Result(i: 2.000000 | j: 1.000000 | i*j: 2.000000): 62.741782 ITER: 0 | i: 4 | j: 2 Result(i: 2.000000 | j: 1.414214 | i*j: 2.828427): 65.570209 ITER: 0 | i: 4 | j: 3 Result(i: 2.000000 | j: 1.732051 | i*j: 3.464102): 69.034310 ITER: 0 | i: 4 | j: 4 Result(i: 2.000000 | j: 2.000000 | i*j: 4.000000): 73.034310 ITER: 0 | i: 4 | j: 5 Result(i: 2.000000 | j: 2.236068 | i*j: 4.472136): 77.506446 ITER: 0 | i: 4 | j: 6 Result(i: 2.000000 | j: 2.449490 | i*j: 4.898979): 82.405426 ITER: 0 | i: 4 | j: 7 Result(i: 2.000000 | j: 2.645751 | i*j: 5.291503): 87.696928 ITER: 0 | i: 4 | j: 8 Result(i: 2.000000 | j: 2.828427 | i*j: 5.656854): 93.353783 ITER: 0 | i: 4 | j: 9 Result(i: 2.000000 | j: 3.000000 | i*j: 6.000000): 99.353783 ITER: 0 | i: 5 | j: 0 Result(i: 2.236068 | j: 0.000000 | i*j: 0.000000): 99.353783 ITER: 0 | i: 5 | j: 1 Result(i: 2.236068 | j: 1.000000 | i*j: 2.236068): 101.589851 ITER: 0 | i: 5 | j: 2 Result(i: 2.236068 | j: 1.414214 | i*j: 3.162278): 104.752128 ITER: 0 | i: 5 | j: 3 Result(i: 2.236068 | j: 1.732051 | i*j: 3.872983): 108.625112 ITER: 0 | i: 5 | j: 4 Result(i: 2.236068 | j: 2.000000 | i*j: 4.472136): 113.097248 ITER: 0 | i: 5 | j: 5 Result(i: 2.236068 | j: 2.236068 | i*j: 5.000000): 118.097248 ITER: 0 | i: 5 | j: 6 Result(i: 2.236068 | j: 2.449490 | i*j: 5.477226): 123.574473 ITER: 0 | i: 5 | j: 7 Result(i: 2.236068 | j: 2.645751 | i*j: 5.916080): 129.490553 ITER: 0 | i: 5 | j: 8 Result(i: 2.236068 | j: 2.828427 | i*j: 6.324555): 135.815108 ITER: 0 | i: 5 | j: 9 Result(i: 2.236068 | j: 3.000000 | i*j: 6.708204): 142.523312 ITER: 0 | i: 6 | j: 0 Result(i: 2.449490 | j: 0.000000 | i*j: 0.000000): 142.523312 ITER: 0 | i: 6 | j: 1 Result(i: 2.449490 | j: 1.000000 | i*j: 2.449490): 144.972802 ITER: 0 | i: 6 | j: 2 Result(i: 2.449490 | j: 1.414214 | i*j: 3.464102): 148.436904 ITER: 0 | i: 6 | j: 3 Result(i: 2.449490 | j: 1.732051 | i*j: 4.242641): 152.679544 ITER: 0 | i: 6 | j: 4 Result(i: 2.449490 | j: 2.000000 | i*j: 4.898979): 157.578524 ITER: 0 | i: 6 | j: 5 Result(i: 2.449490 | j: 2.236068 | i*j: 5.477226): 163.055749 ITER: 0 | i: 6 | j: 6 Result(i: 2.449490 | j: 2.449490 | i*j: 6.000000): 169.055749 ITER: 0 | i: 6 | j: 7 Result(i: 2.449490 | j: 2.645751 | i*j: 6.480741): 175.536490 ITER: 0 | i: 6 | j: 8 Result(i: 2.449490 | j: 2.828427 | i*j: 6.928203): 182.464693 ITER: 0 | i: 6 | j: 9 Result(i: 2.449490 | j: 3.000000 | i*j: 7.348469): 189.813162 ITER: 0 | i: 7 | j: 0 Result(i: 2.645751 | j: 0.000000 | i*j: 0.000000): 189.813162 ITER: 0 | i: 7 | j: 1 Result(i: 2.645751 | j: 1.000000 | i*j: 2.645751): 192.458914 ITER: 0 | i: 7 | j: 2 Result(i: 2.645751 | j: 1.414214 | i*j: 3.741657): 196.200571 ITER: 0 | i: 7 | j: 3 Result(i: 2.645751 | j: 1.732051 | i*j: 4.582576): 200.783147 ITER: 0 | i: 7 | j: 4 Result(i: 2.645751 | j: 2.000000 | i*j: 5.291503): 206.074649 ITER: 0 | i: 7 | j: 5 Result(i: 2.645751 | j: 2.236068 | i*j: 5.916080): 211.990729 ITER: 0 | i: 7 | j: 6 Result(i: 2.645751 | j: 2.449490 | i*j: 6.480741): 218.471470 ITER: 0 | i: 7 | j: 7 Result(i: 2.645751 | j: 2.645751 | i*j: 7.000000): 225.471470 ITER: 0 | i: 7 | j: 8 Result(i: 2.645751 | j: 2.828427 | i*j: 7.483315): 232.954785 ITER: 0 | i: 7 | j: 9 Result(i: 2.645751 | j: 3.000000 | i*j: 7.937254): 240.892039 ITER: 0 | i: 8 | j: 0 Result(i: 2.828427 | j: 0.000000 | i*j: 0.000000): 240.892039 ITER: 0 | i: 8 | j: 1 Result(i: 2.828427 | j: 1.000000 | i*j: 2.828427): 243.720466 ITER: 0 | i: 8 | j: 2 Result(i: 2.828427 | j: 1.414214 | i*j: 4.000000): 247.720466 ITER: 0 | i: 8 | j: 3 Result(i: 2.828427 | j: 1.732051 | i*j: 4.898979): 252.619445 ITER: 0 | i: 8 | j: 4 Result(i: 2.828427 | j: 2.000000 | i*j: 5.656854): 258.276300 ITER: 0 | i: 8 | j: 5 Result(i: 2.828427 | j: 2.236068 | i*j: 6.324555): 264.600855 ITER: 0 | i: 8 | j: 6 Result(i: 2.828427 | j: 2.449490 | i*j: 6.928203): 271.529058 ITER: 0 | i: 8 | j: 7 Result(i: 2.828427 | j: 2.645751 | i*j: 7.483315): 279.012373 ITER: 0 | i: 8 | j: 8 Result(i: 2.828427 | j: 2.828427 | i*j: 8.000000): 287.012373 ITER: 0 | i: 8 | j: 9 Result(i: 2.828427 | j: 3.000000 | i*j: 8.485281): 295.497654 ITER: 0 | i: 9 | j: 0 Result(i: 3.000000 | j: 0.000000 | i*j: 0.000000): 295.497654 ITER: 0 | i: 9 | j: 1 Result(i: 3.000000 | j: 1.000000 | i*j: 3.000000): 298.497654 ITER: 0 | i: 9 | j: 2 Result(i: 3.000000 | j: 1.414214 | i*j: 4.242641): 302.740295 ITER: 0 | i: 9 | j: 3 Result(i: 3.000000 | j: 1.732051 | i*j: 5.196152): 307.936447 ITER: 0 | i: 9 | j: 4 Result(i: 3.000000 | j: 2.000000 | i*j: 6.000000): 313.936447 ITER: 0 | i: 9 | j: 5 Result(i: 3.000000 | j: 2.236068 | i*j: 6.708204): 320.644651 ITER: 0 | i: 9 | j: 6 Result(i: 3.000000 | j: 2.449490 | i*j: 7.348469): 327.993120 ITER: 0 | i: 9 | j: 7 Result(i: 3.000000 | j: 2.645751 | i*j: 7.937254): 335.930374 ITER: 0 | i: 9 | j: 8 Result(i: 3.000000 | j: 2.828427 | i*j: 8.485281): 344.415656 ITER: 0 | i: 9 | j: 9 Result(i: 3.000000 | j: 3.000000 | i*j: 9.000000): 353.415656 Final Result: 353.415656 ITER: 1 | i: 0 | j: 0 Result(i: 0.000000 | j: 0.000000 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 1 Result(i: 0.000000 | j: 1.000000 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 2 Result(i: 0.000000 | j: 1.414214 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 3 Result(i: 0.000000 | j: 1.732051 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 4 Result(i: 0.000000 | j: 2.000000 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 5 Result(i: 0.000000 | j: 2.236068 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 6 Result(i: 0.000000 | j: 2.449490 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 7 Result(i: 0.000000 | j: 2.645751 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 8 Result(i: 0.000000 | j: 2.828427 | i*j: 0.000000): 0.000000 ITER: 1 | i: 0 | j: 9 Result(i: 0.000000 | j: 3.000000 | i*j: 0.000000): 0.000000 ITER: 1 | i: 1 | j: 0 Result(i: 1.000000 | j: 0.000000 | i*j: 0.000000): 0.000000 ITER: 1 | i: 1 | j: 1 Result(i: 1.000000 | j: 1.000000 | i*j: 1.000000): 1.000000 ITER: 1 | i: 1 | j: 2 Result(i: 1.000000 | j: 1.414214 | i*j: 1.414214): 2.414214 ITER: 1 | i: 1 | j: 3 Result(i: 1.000000 | j: 1.732051 | i*j: 1.732051): 4.146264 ITER: 1 | i: 1 | j: 4 Result(i: 1.000000 | j: 2.000000 | i*j: 2.000000): 6.146264 ITER: 1 | i: 1 | j: 5 Result(i: 1.000000 | j: 2.236068 | i*j: 2.236068): 8.382332 ITER: 1 | i: 1 | j: 6 Result(i: 1.000000 | j: 2.449490 | i*j: 2.449490): 10.831822 ITER: 1 | i: 1 | j: 7 Result(i: 1.000000 | j: 2.645751 | i*j: 2.645751): 13.477573 ITER: 1 | i: 1 | j: 8 Result(i: 1.000000 | j: 2.828427 | i*j: 2.828427): 16.306001 ITER: 1 | i: 1 | j: 9 Result(i: 1.000000 | j: 3.000000 | i*j: 3.000000): 19.306001 ITER: 1 | i: 2 | j: 0 Result(i: 1.414214 | j: 0.000000 | i*j: 0.000000): 19.306001 ITER: 1 | i: 2 | j: 1 Result(i: 1.414214 | j: 1.000000 | i*j: 1.414214): 20.720214 ITER: 1 | i: 2 | j: 2 Result(i: 1.414214 | j: 1.414214 | i*j: 2.000000): 22.720214 ITER: 1 | i: 2 | j: 3 Result(i: 1.414214 | j: 1.732051 | i*j: 2.449490): 25.169704 ITER: 1 | i: 2 | j: 4 Result(i: 1.414214 | j: 2.000000 | i*j: 2.828427): 27.998131 ITER: 1 | i: 2 | j: 5 Result(i: 1.414214 | j: 2.236068 | i*j: 3.162278): 31.160409 ITER: 1 | i: 2 | j: 6 Result(i: 1.414214 | j: 2.449490 | i*j: 3.464102): 34.624510 ITER: 1 | i: 2 | j: 7 Result(i: 1.414214 | j: 2.645751 | i*j: 3.741657): 38.366168 ITER: 1 | i: 2 | j: 8 Result(i: 1.414214 | j: 2.828427 | i*j: 4.000000): 42.366168 ITER: 1 | i: 2 | j: 9 Result(i: 1.414214 | j: 3.000000 | i*j: 4.242641): 46.608808 ITER: 1 | i: 3 | j: 0 Result(i: 1.732051 | j: 0.000000 | i*j: 0.000000): 46.608808 ITER: 1 | i: 3 | j: 1 Result(i: 1.732051 | j: 1.000000 | i*j: 1.732051): 48.340859 ITER: 1 | i: 3 | j: 2 Result(i: 1.732051 | j: 1.414214 | i*j: 2.449490): 50.790349 ITER: 1 | i: 3 | j: 3 Result(i: 1.732051 | j: 1.732051 | i*j: 3.000000): 53.790349 ITER: 1 | i: 3 | j: 4 Result(i: 1.732051 | j: 2.000000 | i*j: 3.464102): 57.254450 ITER: 1 | i: 3 | j: 5 Result(i: 1.732051 | j: 2.236068 | i*j: 3.872983): 61.127434 ITER: 1 | i: 3 | j: 6 Result(i: 1.732051 | j: 2.449490 | i*j: 4.242641): 65.370075 ITER: 1 | i: 3 | j: 7 Result(i: 1.732051 | j: 2.645751 | i*j: 4.582576): 69.952650 ITER: 1 | i: 3 | j: 8 Result(i: 1.732051 | j: 2.828427 | i*j: 4.898979): 74.851630 ITER: 1 | i: 3 | j: 9 Result(i: 1.732051 | j: 3.000000 | i*j: 5.196152): 80.047782 ITER: 1 | i: 4 | j: 0 Result(i: 2.000000 | j: 0.000000 | i*j: 0.000000): 80.047782 ITER: 1 | i: 4 | j: 1 Result(i: 2.000000 | j: 1.000000 | i*j: 2.000000): 82.047782 ITER: 1 | i: 4 | j: 2 Result(i: 2.000000 | j: 1.414214 | i*j: 2.828427): 84.876209 ITER: 1 | i: 4 | j: 3 Result(i: 2.000000 | j: 1.732051 | i*j: 3.464102): 88.340311 ITER: 1 | i: 4 | j: 4 Result(i: 2.000000 | j: 2.000000 | i*j: 4.000000): 92.340311 ITER: 1 | i: 4 | j: 5 Result(i: 2.000000 | j: 2.236068 | i*j: 4.472136): 96.812447 ITER: 1 | i: 4 | j: 6 Result(i: 2.000000 | j: 2.449490 | i*j: 4.898979): 101.711426 ITER: 1 | i: 4 | j: 7 Result(i: 2.000000 | j: 2.645751 | i*j: 5.291503): 107.002929 ITER: 1 | i: 4 | j: 8 Result(i: 2.000000 | j: 2.828427 | i*j: 5.656854): 112.659783 ITER: 1 | i: 4 | j: 9 Result(i: 2.000000 | j: 3.000000 | i*j: 6.000000): 118.659783 ITER: 1 | i: 5 | j: 0 Result(i: 2.236068 | j: 0.000000 | i*j: 0.000000): 118.659783 ITER: 1 | i: 5 | j: 1 Result(i: 2.236068 | j: 1.000000 | i*j: 2.236068): 120.895851 ITER: 1 | i: 5 | j: 2 Result(i: 2.236068 | j: 1.414214 | i*j: 3.162278): 124.058129 ITER: 1 | i: 5 | j: 3 Result(i: 2.236068 | j: 1.732051 | i*j: 3.872983): 127.931112 ITER: 1 | i: 5 | j: 4 Result(i: 2.236068 | j: 2.000000 | i*j: 4.472136): 132.403248 ITER: 1 | i: 5 | j: 5 Result(i: 2.236068 | j: 2.236068 | i*j: 5.000000): 137.403248 ITER: 1 | i: 5 | j: 6 Result(i: 2.236068 | j: 2.449490 | i*j: 5.477226): 142.880474 ITER: 1 | i: 5 | j: 7 Result(i: 2.236068 | j: 2.645751 | i*j: 5.916080): 148.796553 ITER: 1 | i: 5 | j: 8 Result(i: 2.236068 | j: 2.828427 | i*j: 6.324555): 155.121109 ITER: 1 | i: 5 | j: 9 Result(i: 2.236068 | j: 3.000000 | i*j: 6.708204): 161.829313 ITER: 1 | i: 6 | j: 0 Result(i: 2.449490 | j: 0.000000 | i*j: 0.000000): 161.829313 ITER: 1 | i: 6 | j: 1 Result(i: 2.449490 | j: 1.000000 | i*j: 2.449490): 164.278802 ITER: 1 | i: 6 | j: 2 Result(i: 2.449490 | j: 1.414214 | i*j: 3.464102): 167.742904 ITER: 1 | i: 6 | j: 3 Result(i: 2.449490 | j: 1.732051 | i*j: 4.242641): 171.985545 ITER: 1 | i: 6 | j: 4 Result(i: 2.449490 | j: 2.000000 | i*j: 4.898979): 176.884524 ITER: 1 | i: 6 | j: 5 Result(i: 2.449490 | j: 2.236068 | i*j: 5.477226): 182.361750 ITER: 1 | i: 6 | j: 6 Result(i: 2.449490 | j: 2.449490 | i*j: 6.000000): 188.361750 ITER: 1 | i: 6 | j: 7 Result(i: 2.449490 | j: 2.645751 | i*j: 6.480741): 194.842491 ITER: 1 | i: 6 | j: 8 Result(i: 2.449490 | j: 2.828427 | i*j: 6.928203): 201.770694 ITER: 1 | i: 6 | j: 9 Result(i: 2.449490 | j: 3.000000 | i*j: 7.348469): 209.119163 ITER: 1 | i: 7 | j: 0 Result(i: 2.645751 | j: 0.000000 | i*j: 0.000000): 209.119163 ITER: 1 | i: 7 | j: 1 Result(i: 2.645751 | j: 1.000000 | i*j: 2.645751): 211.764914 ITER: 1 | i: 7 | j: 2 Result(i: 2.645751 | j: 1.414214 | i*j: 3.741657): 215.506572 ITER: 1 | i: 7 | j: 3 Result(i: 2.645751 | j: 1.732051 | i*j: 4.582576): 220.089147 ITER: 1 | i: 7 | j: 4 Result(i: 2.645751 | j: 2.000000 | i*j: 5.291503): 225.380650 ITER: 1 | i: 7 | j: 5 Result(i: 2.645751 | j: 2.236068 | i*j: 5.916080): 231.296730 ITER: 1 | i: 7 | j: 6 Result(i: 2.645751 | j: 2.449490 | i*j: 6.480741): 237.777470 ITER: 1 | i: 7 | j: 7 Result(i: 2.645751 | j: 2.645751 | i*j: 7.000000): 244.777470 ITER: 1 | i: 7 | j: 8 Result(i: 2.645751 | j: 2.828427 | i*j: 7.483315): 252.260785 ITER: 1 | i: 7 | j: 9 Result(i: 2.645751 | j: 3.000000 | i*j: 7.937254): 260.198039 ITER: 1 | i: 8 | j: 0 Result(i: 2.828427 | j: 0.000000 | i*j: 0.000000): 260.198039 ITER: 1 | i: 8 | j: 1 Result(i: 2.828427 | j: 1.000000 | i*j: 2.828427): 263.026466 ITER: 1 | i: 8 | j: 2 Result(i: 2.828427 | j: 1.414214 | i*j: 4.000000): 267.026466 ITER: 1 | i: 8 | j: 3 Result(i: 2.828427 | j: 1.732051 | i*j: 4.898979): 271.925446 ITER: 1 | i: 8 | j: 4 Result(i: 2.828427 | j: 2.000000 | i*j: 5.656854): 277.582300 ITER: 1 | i: 8 | j: 5 Result(i: 2.828427 | j: 2.236068 | i*j: 6.324555): 283.906855 ITER: 1 | i: 8 | j: 6 Result(i: 2.828427 | j: 2.449490 | i*j: 6.928203): 290.835059 ITER: 1 | i: 8 | j: 7 Result(i: 2.828427 | j: 2.645751 | i*j: 7.483315): 298.318373 ITER: 1 | i: 8 | j: 8 Result(i: 2.828427 | j: 2.828427 | i*j: 8.000000): 306.318373 ITER: 1 | i: 8 | j: 9 Result(i: 2.828427 | j: 3.000000 | i*j: 8.485281): 314.803655 ITER: 1 | i: 9 | j: 0 Result(i: 3.000000 | j: 0.000000 | i*j: 0.000000): 314.803655 ITER: 1 | i: 9 | j: 1 Result(i: 3.000000 | j: 1.000000 | i*j: 3.000000): 317.803655 ITER: 1 | i: 9 | j: 2 Result(i: 3.000000 | j: 1.414214 | i*j: 4.242641): 322.046295 ITER: 1 | i: 9 | j: 3 Result(i: 3.000000 | j: 1.732051 | i*j: 5.196152): 327.242448 ITER: 1 | i: 9 | j: 4 Result(i: 3.000000 | j: 2.000000 | i*j: 6.000000): 333.242448 ITER: 1 | i: 9 | j: 5 Result(i: 3.000000 | j: 2.236068 | i*j: 6.708204): 339.950652 ITER: 1 | i: 9 | j: 6 Result(i: 3.000000 | j: 2.449490 | i*j: 7.348469): 347.299121 ITER: 1 | i: 9 | j: 7 Result(i: 3.000000 | j: 2.645751 | i*j: 7.937254): 355.236375 ITER: 1 | i: 9 | j: 8 Result(i: 3.000000 | j: 2.828427 | i*j: 8.485281): 363.721656 ITER: 1 | i: 9 | j: 9 Result(i: 3.000000 | j: 3.000000 | i*j: 9.000000): 372.721656 Final Result: 372.721656 As we can see in the following iterations the sqrt(1) as well as the result is set to zero for some reason. ITER: 0 | i: 1 | j: 4 Result(i: 0.000000 | j: 2.000000 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 5 Result(i: 0.000000 | j: 2.236068 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 6 Result(i: 0.000000 | j: 2.449490 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 7 Result(i: 0.000000 | j: 2.645751 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 8 Result(i: 0.000000 | j: 2.828427 | i*j: 0.000000): 0.000000 ITER: 0 | i: 1 | j: 9 Result(i: 0.000000 | j: 3.000000 | i*j: 0.000000): 0.000000 Please help me to resolve the accuracy issue! I think that it will be very useful for gem5 community. To be noticed, I find the correct simulated tick in which the application started in FS (using m5 dumpstats), and I start the --debug-start, but the trace file which is generated is 10x larger than SE mode for the same application. How can I compare them? Thank you in advance! Best regards, Nikos Quoting Νικόλαος Ταμπουρατζής <ntampouratzis@ece.auth.gr>: > Dear Jason, > > I am trying to use --debug-start but in FS mode it is very difficult > to find the tick on which the application is started! > > However, I am writing the following very simple c++ program: > > #include <cmath> > #include <stdio.h> > > int main(){ > > int dim = 4096; > > double result; > > for (int iter = 0; iter < 2; iter++){ > result = 0; > for (int i = 0; i < dim; i++){ > for (int j = 0; j < dim; j++){ > result += sqrt(i) * sqrt(j); > } > } > printf("Result: %lf\n", result); //Result: 30530733453.127449 > } > } > > I cross-compile it using: riscv64-linux-gnu-g++ -static -O3 -o > test_riscv test_riscv.cpp > > > While in X86 (without cross-compilation of course), QEMU-RISCV, > GEM5-SE the result is the same (30530733453.127449), in GEM5-FS the > result is different! In addition, the result is also different > between the 2 iterations. > > Please reproduce the error if you want in order to verify my result. > Ηow can the issue be resolved? > > Thank you in advance! > > Best regards, > Nikos > > > Quoting Jason Lowe-Power <jason@lowepower.com>: > >> Hi Nikos, >> >> You can use --debug-start to start the debugging after some number of >> ticks. Also, I would expect that the difference should come up quickly, so >> no need to run the program to the end. >> >> For the FS mode one, you will want to just start the trace as the >> application starts. This could be a bit of a pain. >> >> I'm not really sure what fundamentally could be different. FS and SE mode >> use the exact same code for executing instructions, so I don't think that's >> the problem. Have you tried running for smaller inputs or just one >> iteration? >> >> Jason >> >> >> >> On Wed, Sep 21, 2022 at 9:04 AM Νικόλαος Ταμπουρατζής < >> ntampouratzis@ece.auth.gr> wrote: >> >>> Dear Bobby, >>> >>> Iam trying to add --debug-flags=Exec (building the gem5 for gem5.opt >>> not for gem5.fast which I had) but the debug traces exceed the 20GB >>> (and it is not finished yet) for less than 1 simulated second. How can >>> I reduce the size of the debug-flags (or set something more specific)? >>> >>> In contrast I build the HPCG benchmark with DHPCG_DEBUG flag. If you >>> want, you can compare these two output files >>> (hpcg20010909T014640_SE_Mode & HPCG-Benchmark_3.1_FS_Mode). As you can >>> see, something goes wrong with the accuracy of calculations in FS mode >>> (benchmark uses double precission). You can find the files here: >>> http://kition.mhl.tuc.gr:8000/d/68d82f3533/ >>> >>> Best regards, >>> Nikos >>> >>> Quoting Jason Lowe-Power <jason@lowepower.com>: >>> >>>> That's quite odd that it works in SE mode but not FS mode! >>>> >>>> I would suggest running with --debug-flags=Exec for both and then >>> perform a >>>> diff to see how they differ. >>>> >>>> Cheers, >>>> Jason >>>> >>>> On Tue, Sep 20, 2022 at 2:45 PM Νικόλαος Ταμπουρατζής < >>>> ntampouratzis@ece.auth.gr> wrote: >>>> >>>>> Dear Bobby, >>>>> >>>>> In QEMU I get the same (correct) results that I get in SE mode >>>>> simulation. I get invalid results in FS simulation (in both >>>>> riscv-fs.py and riscv-ubuntu-run.py). I cannot access real RISCV >>>>> hardware at this moment, however, if you want you may execute my xhpcg >>>>> binary (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) with the >>>>> following configuration: >>>>> >>>>> ./xhpcg --nx=16 --ny=16 --nz=16 --npx=1 --npy=1 --npz=1 --rt=0.1 >>>>> >>>>> Please let me know if you have any updates! >>>>> >>>>> Best regards, >>>>> Nikos >>>>> >>>>> >>>>> Quoting Jason Lowe-Power <jason@lowepower.com>: >>>>> >>>>> > Hi Nikos, >>>>> > >>>>> > I notice you said the following in your original email: >>>>> > >>>>> > In addition, I used the RISCV Ubuntu image >>>>> >> (https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu >>> ), >>>>> >> I installed the gcc compiler, compile it (through qemu) and I get >>>>> >> wrong results too. >>>>> > >>>>> > >>>>> > Is this saying you get the wrong results is QEMU? If so, the bug is in >>>>> GCC >>>>> > or the HPCG workload, not in gem5. If not, I would test in QEMU to >>> make >>>>> > sure the binary works there. Another way you could test to see if the >>>>> > problem is your binary or gem5 would be to run it on real hardware. We >>>>> have >>>>> > access to some RISC-V hardware here at UC Davis, if you don't have >>> access >>>>> > to it. >>>>> > >>>>> > Cheers, >>>>> > Jason >>>>> > >>>>> > On Tue, Sep 20, 2022 at 12:58 AM Νικόλαος Ταμπουρατζής < >>>>> > ntampouratzis@ece.auth.gr> wrote: >>>>> > >>>>> >> Dear Bobby, >>>>> >> >>>>> >> 1) I use the original riscv-fs.py which is provided in the latest >>> gem5 >>>>> >> release. >>>>> >> I run the gem5 once (./build/RISCV/gem5.fast -d ./HPCG_FS_results >>>>> >> ./configs/example/gem5_library/riscv-fs.py) in order to download the >>>>> >> riscv-bootloader-vmlinux-5.10 and riscv-disk-img. >>>>> >> After this I mount the riscv-disk-img (sudo mount -o loop >>>>> >> riscv-disk-img /mnt), put the xhpcg executable and I do the following >>>>> >> changes in riscv-fs.py to boot the riscv-disk-img with executable: >>>>> >> >>>>> >> image = CustomDiskImageResource( >>>>> >> local_path = "/home/cossim/.cache/gem5/riscv-disk-img", >>>>> >> ) >>>>> >> >>>>> >> # Set the Full System workload. >>>>> >> board.set_kernel_disk_workload( >>>>> >> kernel=Resource("riscv-bootloader-vmlinux-5.10"), >>>>> >> disk_image=image, >>>>> >> ) >>>>> >> >>>>> >> Finally, in the gem5/src/python/gem5/components/boards/riscv_board.py >>>>> >> I change the last line to "return ["console=ttyS0", >>>>> >> "root={root_value}", "rw"]" in order to allow the write permissions >>> in >>>>> >> the image. >>>>> >> >>>>> >> >>>>> >> 2) The HPCG benchmark after some iterations calculates if the results >>>>> >> are valid or not valid. In the case of FS it gives invalid results. >>> As >>>>> >> I see from the results, one (at least) problem is that produces >>>>> >> different results in each HPCG execution (with the same >>> configuration). >>>>> >> >>>>> >> Here is the HPCG output and riscv-fs.py >>>>> >> (http://kition.mhl.tuc.gr:8000/d/68d82f3533/). You may reproduce the >>>>> >> results in the video if you use the xhpcg executable >>>>> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/) >>>>> >> >>>>> >> Please help me in order to solve it! >>>>> >> >>>>> >> Finally, I get invalid results in the HPL benchmark in FS mode too. >>>>> >> >>>>> >> Best regards, >>>>> >> Nikos >>>>> >> >>>>> >> >>>>> >> Quoting Bobby Bruce <bbruce@ucdavis.edu>: >>>>> >> >>>>> >> > I'm going to need a bit more information to help: >>>>> >> > >>>>> >> > 1. In what way have you modified >>>>> >> > ./configs/example/gem5_library/riscv-fs.py? Can you attach the >>> script >>>>> >> here? >>>>> >> > 2. What error are you getting or in what way are the results >>> invalid? >>>>> >> > >>>>> >> > - >>>>> >> > Dr. Bobby R. Bruce >>>>> >> > Room 3050, >>>>> >> > Kemper Hall, UC Davis >>>>> >> > Davis, >>>>> >> > CA, 95616 >>>>> >> > >>>>> >> > web: https://www.bobbybruce.net >>>>> >> > >>>>> >> > >>>>> >> > On Mon, Sep 19, 2022 at 1:43 PM Νικόλαος Ταμπουρατζής < >>>>> >> > ntampouratzis@ece.auth.gr> wrote: >>>>> >> > >>>>> >> >> >>>>> >> >> Dear gem5 community, >>>>> >> >> >>>>> >> >> I have successfully cross-compile the HPCG benchmark for RISCV >>>>> (Serial >>>>> >> >> version, without MPI and OpenMP). While it working properly in >>> gem5 >>>>> SE >>>>> >> >> mode (./build/RISCV/gem5.fast -d ./HPCG_SE_results >>>>> >> >> ./configs/example/se.py -c xhpcg --options '--nx=16 --ny=16 >>> --nz=16 >>>>> >> >> --npx=1 --npy=1 --npz=1 --rt=0.1'), I get invalid results in FS >>>>> >> >> simulation using "./build/RISCV/gem5.fast -d ./HPCG_FS_results >>>>> >> >> ./configs/example/gem5_library/riscv-fs.py" (I mount the riscv >>> image >>>>> >> >> and put it). >>>>> >> >> >>>>> >> >> Can you help me please? >>>>> >> >> >>>>> >> >> In addition, I used the RISCV Ubuntu image >>>>> >> >> ( >>> https://github.com/gem5/gem5-resources/tree/stable/src/riscv-ubuntu >>>>> ), >>>>> >> >> I installed the gcc compiler, compile it (through qemu) and I get >>>>> >> >> wrong results too. >>>>> >> >> >>>>> >> >> Here is the Makefile which I use, the hpcg executable for RISCV >>>>> >> >> (xhpcg), and a video that shows the results >>>>> >> >> (http://kition.mhl.tuc.gr:8000/f/4ca25fdd3c/). >>>>> >> >> >>>>> >> >> P.S. I use the latest gem5 version. >>>>> >> >> >>>>> >> >> Thank you in advance! :) >>>>> >> >> >>>>> >> >> Best regards, >>>>> >> >> Nikos >>>>> >> >> _______________________________________________ >>>>> >> >> gem5-users mailing list -- gem5-users@gem5.org >>>>> >> >> To unsubscribe send an email to gem5-users-leave@gem5.org >>>>> >> >> >>>>> >> >>>>> >> >>>>> >> _______________________________________________ >>>>> >> gem5-users mailing list -- gem5-users@gem5.org >>>>> >> To unsubscribe send an email to gem5-users-leave@gem5.org >>>>> >> >>>>> >>>>> >>>>> _______________________________________________ >>>>> gem5-users mailing list -- gem5-users@gem5.org >>>>> To unsubscribe send an email to gem5-users-leave@gem5.org >>>>> >>> >>> >>> _______________________________________________ >>> gem5-users mailing list -- gem5-users@gem5.org >>> To unsubscribe send an email to gem5-users-leave@gem5.org >>> > > > _______________________________________________ > gem5-users mailing list -- gem5-users@gem5.org > To unsubscribe send an email to gem5-users-leave@gem5.org