LinX - A simple Linpack interface

**Fritz_the_germ** · 08-31-2009, 12:20 PM

Originally Posted by gvansly1

I have not searched the posts for this question.....but I'm getting a "Stopped upon error after 3 m 40 s!"

Is this a vcore, vtt issue or somwthing else? No BSOD or lockups ...

Everytime I had an error I had to increase the nb voltage or sometimes the vtt.
You have to find it out by trial and error

**SimpleTECH** · 09-02-2009, 06:51 PM

Version 0.6.3 has been released.

http://www.youwatched.com/datajay/linx(0.63).7z

Changelog:

removed the easter egg and added an option to extend glass for Windows Vista and Windows 7 users with Aero on
changed (raised) upper limit for 32-bit Linpack to 15500
graph selection is now saved
monitoring data (minimum and maximum values for each monitor) is now appended to the end of text log if graphs were enabled during testing
fixed a possible bug when LinX would stop responding if Linpack quits due to lack of memory
fixed a possible bug when running more than one instance of LinX simultaneously
fixed a possible bug in graphs that would lead to an error message if Everest or Speedfan was shut down during testing
fixed a bug with progressbar not visible on Classic theme

**GigaByte** · 09-02-2009, 08:50 PM

Small bug, when aero is enabled text at the top of the results chart acts weird, see screenshot.

**Dua|ist** · 09-03-2009, 07:31 AM

Originally Posted by SimpleTECH

Version 0.6.3 has been released.

http://www.youwatched.com/datajay/linx(0.63).7z

Changelog:

removed the easter egg and added an option to extend glass for Windows Vista and Windows 7 users with Aero on
changed (raised) upper limit for 32-bit Linpack to 15500
graph selection is now saved
monitoring data (minimum and maximum values for each monitor) is now appended to the end of text log if graphs were enabled during testing
fixed a possible bug when LinX would stop responding if Linpack quits due to lack of memory
fixed a possible bug when running more than one instance of LinX simultaneously
fixed a possible bug in graphs that would lead to an error message if Everest or Speedfan was shut down during testing
fixed a bug with progressbar not visible on Classic theme

He-he, looks like I don't have to do that part of the job, thanks!

Updated 1st post then.

One more notice: If anyone ever had a problem with LinX «not noticing» that testing is finished and keeping running after all tests are complete, give this version a try. I'm still not sure what could have caused that (pretty rare) bug, but I think and hope that now it's gone.

GigaByte, thanks for the report. What OS are you using? Everything's fine on my Win7 x64.

EDIT: There's also a new permanent link for the latest version's installer if anyone needs it: http://dl.getdropbox.com/u/663563/LinX/LinX-setup.exe

**GigaByte** · 09-03-2009, 08:01 AM

Vista Ultimate x64 SP2.

**donmarkoni** · 09-03-2009, 11:30 AM

Originally Posted by GigaByte

Vista Ultimate x64 SP2.

Same system here and same bug when using glass effect.

**slim142** · 09-03-2009, 12:04 PM

so wait what was the eastern egg in 0.6.2?

**Dua|ist** · 09-04-2009, 06:32 AM

Originally Posted by slim142

so wait what was the eastern egg in 0.6.2?

Glass.

GigaByte, donmarkoni
Ok, I'll try to fix that.

**slim142** · 09-04-2009, 09:22 AM

Originally Posted by Dua|ist

Glass.

GigaByte, donmarkoni
Ok, I'll try to fix that.

Oh well I have XP so no point for me

**AceBaran** · 09-12-2009, 07:45 AM

suguestions to better my timings/volts/etc.. are welcome

**nigelke** · 10-03-2009, 05:20 AM

going test this litle baby out

looks very intresting ^^

**leo27** · 10-04-2009, 05:24 PM

i also have the glass bug on vista x64 SP1

**Dua|ist** · 10-05-2009, 08:54 PM

Glass bug is fixed in the next version I hope. I'm a bit busy at the moment, so I count on your patience. There will be the ability to test for a specified number of minutes in the new version and maybe something else. We'll see.

**paul_** · 10-07-2009, 09:52 AM

Awesome, Dual|st, keep up the good work! Life is a busy thing, and I hope you have some free time to work on this at some point. I look forward to the next version!

**KURTZ** · 10-17-2009, 04:28 AM

@ Dual|st: we made many tests with your superb tool, but unfortunately we find out a difference between LINX and PRIME95, right now (for the Lynnfield processors) the better tool to test the stability of our system is Prime, normally if we were stable with Linx @ x (example 1.35) vCC we need to increase that voltage to 1.37/and over with Prime ... i think LINX need some improvements onto the DIMMs sector ...

just my opinion, obviously

**WaterFlex** · 10-17-2009, 04:43 AM

KURTZ
Linpack doesn`t stress DRAM like 24hours running of PRIME and Windows MEMTEST 3.8 do.

**KURTZ** · 10-17-2009, 05:03 AM

Originally Posted by WaterFlex

KURTZ
Linpack doesn`t stress DRAM like 24hours running of PRIME and Windows MEMTEST 3.8 do.

i know that dude, but i saw many 'solid' test with 50 Linx's loops don't pass 5 mins with Prime ...

**Sniper** · 10-17-2009, 10:47 AM

That can go both ways. You can have oc that passes Prime blend/small etc, and fails LinX, and you can have oc that passes LinX and fails Prime. Theres also cases where you pass both LinX and Prime and it will fail in some game or other application.

Typically LinX will fail before Prime if you are on the edge of what your cpu can do in Mhz / temps, Prime (Blend) will often fail before LinX if your on the edge of memory oc.

Both are valuable tools in verifying oc stability.

**Dua|ist** · 10-19-2009, 08:07 AM

I would agree with Sniper, well said, everything's got its good points.
Anyway, I won't be able to do anything with error finding efficiency as it is Linpack that's responsible for that.

**jcool** · 10-21-2009, 10:03 AM

Dua|list, there seems to be some kind of bug with LinX if ran on a Gulftown CPU:

No, that is not a waterchiller, it's not ready yet

For comparison, temps with prime blend:

**WaterFlex** · 10-21-2009, 11:34 AM

jcool
Wow. Thats interesting.

P.S.
omg, where do u find gulftowns ES???

**stasio** · 10-21-2009, 11:43 AM

Originally Posted by WaterFlex

jcool
Wow. Thats interesting.

P.S.
omg, where do u find gulftowns ES???

Look at his sig

**Dua|ist** · 10-22-2009, 08:57 AM

Originally Posted by jcool

Dua|list, there seems to be some kind of bug with LinX if ran on a Gulftown CPU:

No, that is not a waterchiller, it's not ready yet

For comparison, temps with prime blend:

Yep, no good indeed. Could you please a) check if the Linpack_xeon64.exe process is running when this happens and b) possibly try with some very low memory amount (like 256 MB or maybe even lwss) to check if this is consistent. I'd be very grateful if you could do that for me.

**jcool** · 10-22-2009, 09:56 AM

Originally Posted by Dua|ist

Yep, no good indeed. Could you please a) check if the Linpack_xeon64.exe process is running when this happens and b) possibly try with some very low memory amount (like 256 MB or maybe even lwss) to check if this is consistent. I'd be very grateful if you could do that for me.

Yes, Linpack_xeon64.exe is running. And yes, I believe you are correct... using 256MB, it spikes up to 60C max in the 1-2 seconds it takes the CPU to calculate.

Edit: It also goes to 60-61C hottest core with 512MB. 768MB only goes to ~55C. 1GB and up stay at ~52C. Lol?

**Dua|ist** · 10-24-2009, 07:40 AM

jcool
I'd like to say that it might be Linpack's incompatibility with i9, but it might be my fault as well.

Apart from these unusually low temps does everything else seem OK?

Thread: LinX - A simple Linpack interface

Thread Tools

Search Thread

Rate This Thread

Display

Tags for this Thread

Bookmarks

Bookmarks

Posting Permissions