LinX - A simple Linpack interface

[003]

What do you guys think about a spartan interface with all checkboxes moved to the Settings window like this:

Attachment 94644

I just have a feeling that nobody uses them anyway, or at least not too frequently. Not sure only about the x64 option, but again on 32-bit systems it's absolutely useless, while on 64-bit systems, which usually have 4+ GB of RAM, who'd want to run 32-bit Linpack with its memory limitations?

Or simply leave it all the way it is now?

I like that. Makes the interface cleaner and not as intimidating.

[Warboy]

What do you guys think about a spartan interface with all checkboxes moved to the Settings window like this:

Attachment 94644

I just have a feeling that nobody uses them anyway, or at least not too frequently. Not sure only about the x64 option, but again on 32-bit systems it's absolutely useless, while on 64-bit systems, which usually have 4+ GB of RAM, who'd want to run 32-bit Linpack with its memory limitations?

Or simply leave it all the way it is now?

I like it!

[Slyr7.62]

This is a good program . Check the pics and temps: Prime 95 or OCCT gets my E8400 to 42C, LinX max's @ 48C, but I haven't dusted out my PC in a few weeks so who knows. Actually, looking @ 2 other SS's I took, avg temp during LinX was 43-45C for 100% load, I'm not sure about the 48C peak.

Ninja Edit: Going from 2x1GB to 4x1GB ram soon (Ballistix 800 to HyperX 1066)

[Dua|ist]

This is a good program . Check the pics and temps: Prime 95 or OCCT gets my E8400 to 42C, LinX max's @ 48C, but I haven't dusted out my PC in a few weeks so who knows. Actually, looking @ 2 other SS's I took, avg temp during LinX was 43-45C for 100% load, I'm not sure about the 48C peak.

Thanks.
Now overclock it and the difference in temps will be more evident.

[slim142]

I also like it with the spartan interface

[Dua|ist]

OK, thanks, now we'll see if I can find a good compromise.

[INTREPID]

I've been using IntelBurnTest, but I really like LinX. Great work!

[tomb18]

Perhaps someone can help here...
I've just built my Core i7 920 system and I am running at 20x200 1.44 V load QPI=1.3875V and 6G Corsair 1600 8,8,8-24 This passes LinX just fine with 20 passes on all memory. However, if I run Prime95, it fails rather quickly. It seems that in most cases it is the otherway around. Anybody know where I should look next?
I am running Prime96 right now at DDR31400 or so to see if this is a memory issue. But the memory will pass Memtest86 many times at 4GHz and full DDR31600 speeds no problem. Is this a QPI voltage issue? Or the dreaded 20x200 issue? I hope not cause I can't get over Bclk=204 on this system.

Thanks for your help.

[auchkoenig]

Perhaps someone can help here...
I've just built my Core i7 920 system and I am running at 20x200 1.44 V load QPI=1.3875V and 6G Corsair 1600 8,8,8-24 This passes LinX just fine with 20 passes on all memory. However, if I run Prime95, it fails rather quickly. It seems that in most cases it is the otherway around. Anybody know where I should look next?
I am running Prime96 right now at DDR31400 or so to see if this is a memory issue. But the memory will pass Memtest86 many times at 4GHz and full DDR31600 speeds no problem. Is this a QPI voltage issue? Or the dreaded 20x200 issue? I hope not cause I can't get over Bclk=204 on this system.

Thanks for your help.

Which Prime95 test did it fail? Look like a QPI voltage issue to me. Trying raising it to see if it goes away.

[INTREPID]

I don't have a i7 yet, but here is some interesting reading:

Core i7 - Is High VDimm really a Problem?

http://www.anandtech.com/memory/showdoc.aspx?i=3426

http://www.legitreviews.com/article/806/1/

[Dua|ist]

INTREPID, thanks.

Perhaps someone can help here...
I've just built my Core i7 920 system and I am running at 20x200 1.44 V load QPI=1.3875V and 6G Corsair 1600 8,8,8-24 This passes LinX just fine with 20 passes on all memory. However, if I run Prime95, it fails rather quickly. It seems that in most cases it is the otherway around. Anybody know where I should look next?
I am running Prime96 right now at DDR31400 or so to see if this is a memory issue. But the memory will pass Memtest86 many times at 4GHz and full DDR31600 speeds no problem. Is this a QPI voltage issue? Or the dreaded 20x200 issue? I hope not cause I can't get over Bclk=204 on this system.

Thanks for your help.

If the system passes Linpack it shouldn't fail in Prime due to CPU (in a short period of time at least), so this is IMO some memory or NB issue with 1st being more likely. I, however, too am not very familiar with i7 overclocking specifics.

[tomb18]

Thanks for the information. I have noticed that on my way to stability, I had to raise the QPI voltage one step of the time. AT one point, increasing Vcore did nothing, but increasing QPI did result in a stable linx session. Perhaps I need more QPI.
I'am now prining at a lower memory speed but won't know until I get home at the end of the day if this worked.

[tomb18]

Which Prime95 test did it fail? Look like a QPI voltage issue to me. Trying raising it to see if it goes away.

It fails on the first 1024k tests.
I did try changing from 8-8-8-24 1T to 9-9-9-24 2T and it made no difference. Next is QPI voltage. I'm currently at 1.3875

One other item. In the past on my Core 2 Quad platform, stability testing with |Linx usually ended up with a error in the residuals. In Prime95, the error was a stopped thread. Now, I never get these errors only BSOD's. Any idea of the difference? Does Linx stress other areas beyond the CPU and RAM?

[jcool]

Definitely low VTT voltage ("QPI voltage" is CPU VTT really) or failing ram. Check your ram with memtest first, then move on to LinX/prime.

[Dua|ist]

Now, I never get these errors only BSOD's. Any idea of the difference? Does Linx stress other areas beyond the CPU and RAM?

Well, the only reason I see is the new architecture of Nehalems with lots of new stuff moved to the chip.

I have also wondered why people report only BSODs in Linpack on i7. It seems that the very core itself isn't the reason for BSOD. IIRC nehalem's computational core is similar to the one of Conroe/Penryn. If we don't see any non-matching resuduals prior to BSOD like on Core 2s then maybe we can conclude that it is not the cores that are holding Nehalems back but something from that uncore (which is pretty big on chip space- and trainsistor-wise AFAIK)? Only an assumption, I may be totally wrong here.

Maybe i5 without QPI would be even better?

[tomb18]

Yes it is interesting that there are no residual errors. However, the uncore is to a large degree memory logic (l3 cache, memory controller, and QPI). Now on my previous Core 2 Quad system, I almost always had residual errors and I knew this was from my memory since I was running 8G under high clocks.
If I lowered the clocks it would pass. Never a BSOD.
Now, if there is any parallel with the Nehalem, one would think that if the uncore was failing then you would get the same type of failure of residuals.

Now perhaps the BSOD's are a hint? I get 0x00000101 'A Clock interrupt was not received on a secondary processor within the
allocated time interval'.
"
This is probably just a general error or it perhaps could be related to the memory controller which I presume is a "processor" of sorts.

Anyways, I solved my problem as I mentioned previously...I bought a 940
Interestingly, again it was linx stable but not prime95. Upping the voltage 1 notch from auto values fixed the problem even at lowered memory latencies. It seems that it may be a cpu issue afterall...Are you sure that linx doesn't stress memory more than prime 95?

[Dua|ist]

No, I'm not. It is just based on people's experience. LinX does stress memory and detect serious memory errors but Prime and especially MemTest are believed to be much more sensitive to memory instability.

Residual errors are indication of CPU instability too. When the wrong residual is close to others (correct ones) it is normally a CPU fault. Big difference in residuals often indicates memory or NB issues. At least this is what was observed on Core 2s.

As a side note, 64-bit Linpack often causes BSODs on my E8500. Where 32-bit version always has wrong residuals, 64-bit one would sometimes BSOD.

[tomb18]

Hmmm I seem to remember a thread in the Extreme Bandwidth section where "Cronos" was pushing linpack hard. I think he thought that it was mostly memory stressing.
Anyways, very good tool you've made!

[Dua|ist]

Nothing really new, just cosmetic changes and some small fixes: LinX 0.5.7

Changes:
- slight UI redesign: some options moved to the Settings window
- added a better About window

This is the "new" look:

[Dua|ist]

Hmmm I seem to remember a thread in the Extreme Bandwidth section where "Cronos" was pushing linpack hard. I think he thought that it was mostly memory stressing.
Anyways, very good tool you've made!

Cronos is Linpack's pioneer. It was him who promoted it and made it popular. Haven't seen his posts for a long while now though. I actually only remember him saying that Linpack stresses CPU very hard (There was a Linpack thread in Intel section too ).
Thanks!

[jcool]

I liked 0.5.6 better.. oh well, I'll just stick with it for now

[ReverendMaynard]

No, I'm not. It is just based on people's experience. LinX does stress memory and detect serious memory errors but Prime and especially MemTest are believed to be much more sensitive to memory instability.

People experience varies Dualist, but I'll guarantee you mate, Linx using all available memory with large problem sizes is wicked on the memory controller. We're even using it in our warranty department We've seen memtest86 pass memory that fails in Linx repeatedly.

Awesome tool

[Tekxpert]

So whats the verdict with regards to best memory stress tool, Linpack code, Prime 95 or Memtest??

I've been using non-gui Linpack for many months and never had an issue with my system OC. Don't see the point in running its competitors.

[Dua|ist]

So whats the verdict with regards to best memory stress tool, Linpack code, Prime 95 or Memtest??

I've been using non-gui Linpack for many months and never had an issue with my system OC. Don't see the point in running its competitors.

Maybe you're right. Some people say Linpack is enough for both RAM and CPU testing, some prefer running additional, more specific tests like Memtest. If I pass Linpack at full or near full memory I'm usually stable in MemTest too, but I still prefer to run it just to be sure. This might be different on different rigs/processors/chipsets, etc., so it's better to decide for yourself.

It is of course good to have an all-in-one testing program. And even if Linpack isn't the best memory checking tool it is definitely not the worst one either.

Maybe we'll hear some other opinions here.

[zlojack]

I run LinX, Prime95, HCI Memtest and Memtest86+.

Once I get a certain amount of stability in those plus crunching for a couple of days, then I deem my rig stable.