User Activity

  • Posted a comment on discussion Open Discussion on 7-max

    Maybe there are some "Branch Prediction Optimization": https://wccftech.com/amd-branch-prediction-optimization-ryzen-9000-7000-cpus-available-windows-11-23h2/ https://wccftech.com/amd-ryzen-9000-gaming-performance-update-revised-testing-parity-intel-14th-gen-cpus-optimized-branch-prediction-boost/

  • Modified a comment on discussion Open Discussion on 7-max

    I took some time today to resolve the compilation issues in Visual Studio 2022. After running some tests, I found that the results for 2200 and 1400 were not significantly different(with Windows KB5041587 update). btw, Which tests overflow the uop-cache?

  • Modified a comment on discussion Open Discussion on 7-max

    I took some time today to resolve the compilation issues in Visual Studio 2022. After running some tests, I found that the results for 2200 and 1400 were not significantly different. btw, Which tests overflow the uop-cache?

  • Posted a comment on discussion Open Discussion on 7-max

    I took some time today to resolve the compilation issues in Visual Studio 2022. After running some tests, I found that the results for 2200 and 1400 were not significantly different. btw, Which tests overflow the uop-cache?

  • Posted a comment on discussion Open Discussion on 7-max

    ********************************************************************** ** Visual Studio 2017 Developer Command Prompt v15.9.64 ** Copyright (c) 2017 Microsoft Corporation ********************************************************************** [vcvarsall.bat] Environment initialized for: 'x86' C:\Program Files (x86)\Microsoft Visual Studio\2017\Community>cd D:\backup\benchmark\7bench2200\7bench2200-src\CPP\Utils\CPUTest\MemLat C:\Program Files (x86)\Microsoft Visual Studio\2017\Community>d: D:\backup\benchmark\7bench2200\7bench2200-src\CPP\Utils\CPUTest\MemLat>nmake...

  • Posted a comment on discussion Open Discussion on 7-max

    Thanks for your reply. I download the version 2200, but I don't know how to build it. I have try nmake -f CPP/Build.mak, but get error: nmake -f build.mak Microsoft (R) Program Maintenance Utility Version 14.40.33813.0 Copyright (C) Microsoft Corporation. All rights reserved. link -nologo -OPT:REF -OPT:ICF /LARGEADDRESSAWARE /FIXED:NO -out:o\ oleaut32.lib ole32.lib user32.lib advapi32.lib shell32.lib LINK : fatal error LNK1104: cannot open file 'o\' NMAKE : fatal error U1077: 'link -nologo -OPT:REF...

  • Posted a comment on discussion Open Discussion on 7-max

    perf stat -C 0 -e cpu_core/instructions/,cpu_core/branch/,cpu_core/branch-miss/,cpu_core/l2_request.all/ taskset -c 0 ./pipelen r2 PipeLen64 14.00 : Igor Pavlov : Public domain : 2014-01-04 r2 Branches 0 1 0-1 Random Len1 Len2 32 4.39 24.52 13.98 12.91 -3.09 -2.14 64 7.56 28.58 17.79 18.64 1.14 1.69 128 9.24 25.25 19.77 19.35 4.21 -0.83 256 9.96 26.09 17.05 17.45 -1.14 0.80 512 10.30 26.49 17.42 16.84 -3.11 -1.16 1-K 10.47 26.69 17.59 17.95 -1.26 0.72 2-K 8.70 26.78 17.67 17.92 0.36 0.50 4-K 8.73...

  • Posted a comment on discussion Open Discussion on 7-max

    https://www.7-cpu.com/cpu/Zen3.html 4 KB pages mode (64-bit, Linux) Data TLB L1: 64 items (about 800 KB of memory). ?-assoc. Miss penalty = ? cycles. Parallel miss: ? cycle per access Data TLB L2: 2048 items. 8-way. Miss penalty = ? cycles. Parallel miss: ? cycles per access (read from L3) Size Latency Increase Description 32 K 4 64 K 8 4 + 8 (L2) 128 K 10 2 256 K 11 1 512 K 21 10 + 7 (L1 TLB miss) 1 M 36 15 + 35 (L3) 2 M 46 10 4 M 51 5 8 M 53 2 My 5800X is 3.96 cycle in 64K Column "12", not 8 cycle,...

View All

Personal Data

Username:
edisonchan
Joined:
2012-12-28 01:51:32

Projects

  • No projects to display.

Personal Tools