Prime95 version 29.2

[Prime95]

Originally Posted by tului

You need any specific benchmarks you want me to run on my Ryzen 1800X now that I have DDR4-3200 available?

No benchmarks are needed. By default prime95 should now be selecting good FFT implementations for Ryzen. Specifically, the 4M FFT should now yield about the same throughput as Intel's 4-core chips.

[Prime95]

Build 3 is now available.

I'm still seeing cases (this time my Mac) where the program must exit and restart before new benchmark data is used in FFT selection.

[Ducho]

Download link does not work for me. :(

[db597]

I ran a quick benchmark using Build 3 on my Ryzen 1700 and got the following results.
Didn't want the benchmark to take too long, so I selected these settings:

Number of CPU cores to benchmark: 1
Number of workers to benchmark: 1
Time to run each benchmark (in seconds): 10

Not sure if it's a bug. The speed of these timings look more like 8 core / 1 worker results rather than 1 core 1 worker. Maybe I'm mistaken, but can you guys have a look too?

AMD Ryzen 7 1700 Eight-Core Processor          
CPU speed: 3318.73 MHz, 8 hyperthreaded cores
CPU features: 3DNow! Prefetch, SSE, SSE2, SSE4, AVX, AVX2, FMA
L1 cache size: 32 KB
L2 cache size: 512 KB, L3 cache size: 16 MB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
L1 TLBS: 64
L2 TLBS: 1536
Machine topology as determined by hwloc library:
 Machine#0 (total=10379372KB, Backend=Windows, hwlocVersion=1.11.6, ProcessName=prime95.exe)
  NUMANode#0 (local=10379372KB, total=10379372KB)
    Package#0 (CPUVendor=AuthenticAMD, CPUFamilyNumber=23, CPUModelNumber=1, CPUModel="AMD Ryzen 7 1700 Eight-Core Processor          ", CPUStepping=1)
      L3 (size=8192KB, linesize=64, ways=16, Inclusive=0)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x00000003)
              PU#0 (cpuset: 0x00000001)
              PU#1 (cpuset: 0x00000002)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x0000000c)
              PU#2 (cpuset: 0x00000004)
              PU#3 (cpuset: 0x00000008)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x00000030)
              PU#4 (cpuset: 0x00000010)
              PU#5 (cpuset: 0x00000020)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x000000c0)
              PU#6 (cpuset: 0x00000040)
              PU#7 (cpuset: 0x00000080)
      L3 (size=8192KB, linesize=64, ways=16, Inclusive=0)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x00000300)
              PU#8 (cpuset: 0x00000100)
              PU#9 (cpuset: 0x00000200)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x00000c00)
              PU#10 (cpuset: 0x00000400)
              PU#11 (cpuset: 0x00000800)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x00003000)
              PU#12 (cpuset: 0x00001000)
              PU#13 (cpuset: 0x00002000)
        L2 (size=512KB, linesize=64, ways=8, Inclusive=1)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x0000c000)
              PU#14 (cpuset: 0x00004000)
              PU#15 (cpuset: 0x00008000)
Prime95 64-bit version 29.2, RdtscTiming=1
Timings for 2048K FFT length (1 core, 1 worker):  2.32 ms.  Throughput: 431.72 iter/sec.
Timings for 2560K FFT length (1 core, 1 worker):  3.18 ms.  Throughput: 314.64 iter/sec.
Timings for 3072K FFT length (1 core, 1 worker):  4.07 ms.  Throughput: 245.43 iter/sec.
Timings for 3584K FFT length (1 core, 1 worker):  4.93 ms.  Throughput: 202.64 iter/sec.
Timings for 4096K FFT length (1 core, 1 worker):  5.77 ms.  Throughput: 173.37 iter/sec.
Timings for 5120K FFT length (1 core, 1 worker):  7.39 ms.  Throughput: 135.36 iter/sec.
Timings for 6144K FFT length (1 core, 1 worker):  8.94 ms.  Throughput: 111.89 iter/sec.
Timings for 7168K FFT length (1 core, 1 worker): 10.51 ms.  Throughput: 95.12 iter/sec.
Timings for 8192K FFT length (1 core, 1 worker): 12.12 ms.  Throughput: 82.53 iter/sec.

[Ducho]

Originally Posted by Ducho

Download link does not work for me. :(

It works from workplace but not from home.
I'll contact my ISP about that.

P.S. If it's possible to edit a post I just missed the button. :D

[Prime95]

Originally Posted by db597

Not sure if it's a bug. The speed of these timings look more like 8 core / 1 worker results rather than 1 core 1 worker.

That is a bug. I'll work on a fix.

[pepi37]

Link for build 3 please. On first message link is still for version2

[storm5510]

Originally Posted by Prime95

Build 3 is now available...

I don't know what was going on yesterday with 29.2 B3. Windows 10 smartscreen would not allow it to load. Today, no problems at all. Probably had something to do with a system shutdown and restart.

I'm running a LL doublecheck now. Output is consistent with 29.1 B14.

[Prime95]

Originally Posted by pepi37

Link for build 3 please. On first message link is still for version2

Done

[db597]

Another small thing to fix if you have time... This occurs when I'm doing the Torture Test on my 16 thread Ryzen, and mouse over the green icon in the windows system tray. It shows:

Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Te

i.e. It gets truncated on the 13th one, and threads 14/15/16 don't have a "Self-Test" status text in the the pop-up. Not sure if it's P95's limitation or there are too many characters for the Windows system tray to handle. If you count the number of characters it's 127... so there's could be a limit due to the way the variable has been initialised as a short string.

[aurashift]

Originally Posted by db597

Another small thing to fix if you have time... This occurs when I'm doing the Torture Test on my 16 thread Ryzen, and mouse over the green icon in the windows system tray. It shows:

Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Test
Self-Te

i.e. It gets truncated on the 13th one, and threads 14/15/16 don't have a "Self-Test" status text in the the pop-up. Not sure if it's P95's limitation or there are too many characters for the Windows system tray to handle. If you count the number of characters it's 127... so there's could be a limit due to the way the variable has been initialised as a short string.

ive seen something like this on my big ol honkers too.