mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Software (https://www.mersenneforum.org/forumdisplay.php?f=10)
-   -   Prime95 version 29.6/29.7/29.8 (https://www.mersenneforum.org/showthread.php?t=24094)

kriesel 2019-02-18 23:18

throughput benchmark on the infamous i7-8750H / Win 10
 
V29.6b1 x64, set to "1-3,6" workers, with and without hyperthreading, but not all possibilities, 1024k-32768k, completed in about 2 hours, no stall.

Prime95 2019-02-19 18:46

Build 2 is out for Linux and Windows 64-bit

harlee 2019-02-19 19:38

I'm doing P-1 testing and I've noticed that the results for the B.S. extension (i.e.; E=12) isn't being reported by mersenne.org. Logs shows that the B.S. is being sent to the server.

Prime95 2019-02-19 20:07

[QUOTE=harlee;508942]I'm doing P-1 testing and I've noticed that the results for the B.S. extension (i.e.; E=12) isn't being reported by mersenne.org. Logs shows that the B.S. is being sent to the server.[/QUOTE]

Can you give Aaron an exponent to look at? I'm guessing he is parsing the new JSON format to create the web page.

harlee 2019-02-19 20:45

Sorry, forgot to include the exponents as I was being call away. Here are some exponents from two different systems (hardware and OS):

5337463, 5337443, 5337373
18506447, 18506309, 18506269

from prime.log:

Sending result to server: UID: harlee/i5-5250U_1600, M18506447 completed P-1, B1=330000, B2=8745000, E=12, Wh10: <snip>, AID: <snip>

GP2 2019-02-20 00:15

[QUOTE=Prime95;508938]Build 2 is out for Linux and Windows 64-bit[/QUOTE]

The problem I reported for PRP testing of Wagstaff numbers in build 1 is no longer happening.

kriesel 2019-02-20 01:20

[QUOTE=Prime95;508938]Build 2 is out for Linux and Windows 64-bit[/QUOTE]
What does build 2 address?

Prime95 2019-02-20 02:30

[QUOTE=kriesel;508959]What does build 2 address?[/QUOTE]

see [url]https://www.mersenneforum.org/showpost.php?p=508842&postcount=2[/url]

kracker 2019-02-20 03:12

Running the torture test with Small FFTs and only Disable AVX checked, gives me a no FFT lengths available. (29.6 build 2)

Prime95 2019-02-20 03:51

[QUOTE=kracker;508965]Running the torture test with Small FFTs and only Disable AVX checked, gives me a no FFT lengths available. (29.6 build 2)[/QUOTE]

I need to know your CPU architecture as reported in the main window at startup. I'll need caches sizes as reported by CPU dialog box. How many torture test threads were you running? What were the FFT sizes shown in the greyed area of the torture test dialog box?

Flammo 2019-02-20 22:28

[QUOTE=Prime95;508968]I need to know your CPU architecture as reported in the main window at startup. I'll need caches sizes as reported by CPU dialog box. How many torture test threads were you running? What were the FFT sizes shown in the greyed area of the torture test dialog box?[/QUOTE]

Hi, I can replicate the same matter in much the same way as kracken.

CPU Info:
Intel(R) Xeon(R) CPI E3-1225 v5 @ 3.30 GHz
CPU speed 3183,73 MHz, 4 cores
CPU features: Prefetchw, SSE, SSE2, SSE4, AVX, AVX2, FMA
L1 cache size: 4x32 KB, L2 cache size: 4x256 KB, L3 cache size: 8 MB

To replicate the described by kracker, I open Torture Test:
Number of torture tests threads to run: 4
Small FFTs
Torture test settings (area all greyed-out, but showing the following in greyed-out:)
Min FFT size: 42
Max FFT size: 682
Run FFTs in-place: checked
Memory to use: 0
Time to run each FFT: 3 minutes

Run a weaker torture test:
Disable AVX-512: greyed-out
Disable AVX: checked
Disable AVX2 (fused multiple-add): unchecked
Disable SSE2: unchecked

The error message is "No FFT lengths available in teh range specified. Worker stopped."

If I repeat the above, going to Torture Test menu, I notice it is now selected as Blend mode. It didn't save my settings (probably by design). I then reselect Small FFTs, without disabling AVX, it works fine, starting at FMA3 384K FFT length.

Using other torture test types, I find everything works fine with smallest FFTs, disabling the three available AVX/SSE2 options, one at a time.

Disabling AVX with Largest FFTs results in "No FFT lengths available in the range specified". But disabling either AVX2 or SSE2 works fine with the Largest FFTs.

Hope that helps.


All times are UTC. The time now is 20:54.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.