Open wwwuxy opened 1 week ago
Try running without batch-mode=1
. That way the questasim GUI will open and you can debug better. Check the compiler/compiled code to see if there is a missmatch with the PERCIVAL/Big-PERCIVAL versions. With Big-PERCIVAL, if you are using posit64 you will need to run the posit64_testsuite_llvm
instead of posit_testsuite_llvm
.
Can you tell me the relevant configuration(compiler/compiled) requirements ? I couldn't find them in the readme. Thank you very much~ And I successfully ran it on Ubuntu 20.04 before, but now it fails on 22.04. Is there any correlation?
If it worked before, then the compiler should be the same. RV64 from here: https://github.com/artecs-group/llvm-xposit
Then with Big-PERCIVAL you have either posit32 or posit64. You can select that here: https://github.com/artecs-group/PERCIVAL/blob/fd1d154f409cb52c35f578983247411905b34779/core/include/ariane_pkg.sv#L189
For posit64 you have to compile and run posit64_testsuite_llvm
and for posit32 you have to compile and run posit32_testsuite_llvm
.
Have you tried running without batch-mode=1
and debugged using the questasim GUI?
I am not aware if there have to be any changes when using this in Ubuntu 22.04 instead of 20.04. If you find any, please let me know so I can update in the README.
thanks,i compile and run posit64_testsuite_llvm
,and get this fault, can you give me some suggestions? thank you very much~
!
What was the previous error that you solved?
For this, what version of questasim are you using? Also, you may want to try updating the riscv-dbg subrepo. It doesn't seem to be an issue with the posit part of PERCIVAL.
Thanks for your reply~ my Questasim version is 23.1.1,the download URL is https://www.intel.com/content/www/us/en/software-kit/825244/intel-quartus-prime-standard-edition-design-software-version-23-1-1-for-linux.html,the previous error is issu #5
Thanks for opening the other issue, please do that when you encounter a problem and find the solution so I can update the README.
I haven't tried that version of Questasim (I have an older one). I would try updating the riscv-dbg subrepo and meanwhile I ping some of my colleagues to see if they encountered the same issue @albertodbg @MarioInf-UCM
thanks,I use questasim2021 to run make sim elf-bin=posit64_testsuite_llvm.elf QUESTASIM_FLAGS="-novopt -suppress 12110" batch-mode=1
,then meet the same fault, I think the version is no matter
I have not experienced such an error. I am using Questasim 2023.4, but it is Siemens Questasim, not Intel's. BTW, could it be that your license does not allow to run more than 10^7 cycles? I remember in the past that somehting similar happened with the student versions of Modelsim.
Thanks for your reply, I get license fromhttps://licensing.intel.com/psg/s/sales-signup-evaluationlicenses?language=en_US
, In fact, I am a student now,if i run make sim elf-bin=$RISCV/riscv64-unknown-elf/bin/pk target-options=posit64_testsuite_llvm.elf batch-mode=1
, it will operated for 12 hours...As the picture of the title shows
Have you tried updating the riscv-dbg subrepo as I suggested? The error seems to be there, not in PERCIVAL/CVA6
yean,I updating the riscv-dbg,but meet other 4 faults..
Maybe some more things have to be adapted, I'm not sure about that. For now, we haven't replicated this using our setup. I can point you to a similar issue that occurred some time ago: https://github.com/openhwgroup/cva6/issues/842 https://github.com/openhwgroup/cva6/issues/800 https://github.com/pulp-platform/riscv-dbg/pull/102
Thank you for your help~By the way,your work is fantastic~~
Thanks for your words. Let's see if you can get PERCIVAL up and running!
when i run ‘LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libstdc++.so.6 make sim elf-bin=$RISCV/riscv64-unknown-elf/bin/pk target-options=posit_testsuite_llvm.elf batch-mode=1’,I have been stuck in “# Time: 250 ns while running the test Iteration: 2 Instances:/ariane_tb/dut/i'ariane/iucva6/ex_stage_i/pa_gen/pa_top_i/pa_gen/genblk3/genblk3/paul64_st_i/PositEncoder " Here,and operated for 12 hours,i think is something wrong. my chip is R7-9700X.