Prime95 crash with Hyper-threading enabled

Laurent · 2012-08-04, 16:36

Hi,

Every time I enable Hyper-threading, Prime95 application crash on my Windows XP (32-bit) platform (even at stock frequencies) when it is running “Self-test 3200K”.
I got a Windows error: “Event Id : 1000”
Description (Sorry, this is a French version of Windows XP) : Application défaillante prime95.exe, version 27.7.1.0, module défaillant prime95.exe, version 27.7.1.0, adresse de défaillance 0x000510c3.

Prime95 always crash when running 3200K test and at the same 0x000510c3 address.

Without Hyper-threading (and running 4 threads) Prime95 is stable (> 48h) even with my CPU overclocked at 4500MHz.

My Linux 64-bit version of Prime95 is stable (> 48h) with Hyper-threading enabled (and running 8 threads), with CPU overclocked at 4400MHz.

My rig configuration is :
Intel Core i7-2600K (Sandy Bridge) 3400MHz
My RAM is OK (tested with Memtest and Linux version of Prime95)
Windows XP (32-bit) SP 3

Prime95 version is 27.7

What’s wrong with Prime95 3200K test and Hyper-threading ?

Dubslow · 2012-08-04, 18:55

Hmm... there have been reports of weird errors, though not like this for v27.

At a first glance, v27.7 is up to build 4 (27.7.4, see second post in thread).

Does Prime95 itself print any errors? What would make the most sense is something like "ERROR: ROUND OFF (244419.1983) > 0.40".

On the other hand, if Prime95 doesn't print any errors of its own, then I'm not sure what. It could be that Winbloze is overreacting to some perfectly normal occurrence...

Laurent · 2012-08-05, 06:24

Hi,

Prime95 does not print any error in its console. It’s a Windows error pop-up (a sort of “seg fault”, also logged in Windows event log) and it seems that Windows system kills some Prime95 worker threads but the other threads continue their work ! Clicking on the Windows error pop-up kills the whole Prime95 process.

I’m not sure if this is hardware related (as it works fine under Linux OS), maybe it is my Windows XP installation or a device driver which doesn’t like Hyper-threading.

I will try the new 27.7 build 4 version.

Thanks.

Laurent · 2012-08-07, 05:10

Hi,

Can someone let the 3200K FFT test running in a loop for some time with a lot of memory used on a Sandy Bridge with Hyper-threading enabled ?

I will run some Memtest passes on my computer to be sure it is not memory related.

Dubslow · 2012-08-07, 05:20

I'd volunteer, mais malheureusement I only use Linux

Perhaps you might get a larger response if you posted in the thread linked above, which is dedicated to discussing the program itself.

2012-08-04, 16:36	#1
Laurent Aug 2012 7₁₀ Posts	Prime95 crash with Hyper-threading enabled Hi, Every time I enable Hyper-threading, Prime95 application crash on my Windows XP (32-bit) platform (even at stock frequencies) when it is running “Self-test 3200K”. I got a Windows error: “Event Id : 1000” Description (Sorry, this is a French version of Windows XP) : Application défaillante prime95.exe, version 27.7.1.0, module défaillant prime95.exe, version 27.7.1.0, adresse de défaillance 0x000510c3. Prime95 always crash when running 3200K test and at the same 0x000510c3 address. Without Hyper-threading (and running 4 threads) Prime95 is stable (> 48h) even with my CPU overclocked at 4500MHz. My Linux 64-bit version of Prime95 is stable (> 48h) with Hyper-threading enabled (and running 8 threads), with CPU overclocked at 4400MHz. My rig configuration is : Intel Core i7-2600K (Sandy Bridge) 3400MHz My RAM is OK (tested with Memtest and Linux version of Prime95) Windows XP (32-bit) SP 3 Prime95 version is 27.7 What’s wrong with Prime95 3200K test and Hyper-threading ?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Prime95 v28.9 crash on my old PC	MaroX	Software	7	2017-02-25 23:14
How much will hyper threading help	primecrusader	Hardware	7	2016-09-16 01:51
did Prime95 crash while I was gone?	ixfd64	Software	1	2007-02-04 16:36
Prime95 crash	ColdFury	Software	7	2003-10-29 18:39
Running prime95 and NFSNET together on a HT enabled system	TauCeti	NFSNET Discussion	1	2003-07-02 16:26

2012-08-04, 18:55	#2
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 1C35₁₆ Posts	Hmm... there have been reports of weird errors, though not like this for v27. At a first glance, v27.7 is up to build 4 (27.7.4, see second post in thread). Does Prime95 itself print any errors? What would make the most sense is something like "ERROR: ROUND OFF (244419.1983) > 0.40". On the other hand, if Prime95 doesn't print any errors of its own, then I'm not sure what. It could be that Winbloze is overreacting to some perfectly normal occurrence...

2012-08-05, 06:24	#3
Laurent Aug 2012 7 Posts	Hi, Prime95 does not print any error in its console. It’s a Windows error pop-up (a sort of “seg fault”, also logged in Windows event log) and it seems that Windows system kills some Prime95 worker threads but the other threads continue their work ! Clicking on the Windows error pop-up kills the whole Prime95 process. I’m not sure if this is hardware related (as it works fine under Linux OS), maybe it is my Windows XP installation or a device driver which doesn’t like Hyper-threading. I will try the new 27.7 build 4 version. Thanks.

2012-08-07, 05:10	#4
Laurent Aug 2012 7 Posts	Hi, Can someone let the 3200K FFT test running in a loop for some time with a lot of memory used on a Sandy Bridge with Hyper-threading enabled ? I will run some Memtest passes on my computer to be sure it is not memory related.

2012-08-07, 05:20	#5
Dubslow Basketry That Evening! "Bunslow the Bold" Jun 2011 40<A<43 -89<O<-88 3·29·83 Posts	I'd volunteer, mais malheureusement I only use Linux Perhaps you might get a larger response if you posted in the thread linked above, which is dedicated to discussing the program itself.