**The LINPACK Thread**

Status
Not open for further replies.
Re: **The Official TF LINPACK Thread**

MWSnap03712-06-09.jpg

I'm aware of my bsod issue. testing w\ mem86 later. this is using the 10% overclock AI tuning feature in bios. At stock 3.0, I passed with about the same as my failing Gflops pic earlier. Wafflehammer - it might be due to some of the settings. I had to set OS to 3 and recieved a little different Gflops when I used different Problem size. I did about 10 tests. I tried to pull the settings screen w\ this one into paint.net, layer, and show all at once. no joy.
 
Re: **The Official TF LINPACK Thread**

I might run it again next time it's cool, but I can't say that I really like LinX for anything but Stress Testing. From what I've seen, after playing around with it this morning, there are 2 many variables in the Program. I had a hard time just trying to get this version (0.6.4) to run in 32-bit mode. I keep getting a "stopped unexpectantly (not enough memory)" error with this version.
linxstopr.jpg


I have an older version of LinX (0.6.2) and I don't get any errors running it, but I also get a lower score and disabling HT doesn't seem to matter.
linx62.jpg


Personally, in this day and age, I don't see why were running this in 32-bit mode. The new version of LinX (0.6.4) runs perfectly in 64-bit mode (for me) and the results seem to be closer together with or without HT. But 3 runs do take along time and I'm not thrilled about running my system that hard that often. I can game all day long and never get this hot.

linx64644.jpg

linx64648.jpg


And I don't think the AMD users are going to see a difference/improvement compared to the SuperPi Results
 
Re: **The Official TF LINPACK Thread**

I had the same memory issue as well. Was kind of a pain. It defiantly turns up the heat no doubt.

Still waiting for an AMD PII score. I went easy on you guys. 4.0, HT on and no tweaking.
 
Re: **The Official TF LINPACK Thread**

41.4365 GFlops | Philbar71 | Core 2 Quad Q9300 @ 3.5GHz

pwnagej.png


Whoop. pretty good. this is at my daily clock. i'll do a run at max a little later :devil:
 
Re: **The Official TF LINPACK Thread**

The stress test aspect is one of the reasons I like using it as a benchmark, it forces people to use somewhat stable clocks and adequate cooling.

I might run it again next time it's cool, but I can't say that I really like LinX for anything but Stress Testing. From what I've seen, after playing around with it this morning, there are 2 many variables in the Program. I had a hard time just trying to get this version (0.6.4) to run in 32-bit mode. I keep getting a "stopped unexpectantly (not enough memory)" error with this version.

I agree that the Linx interface is more complicated then it needs to be for our purposes but I don't know of any other front ends for linpack that are better.
 
Re: **The Official TF LINPACK Thread**

If you'll look thru the results posted you'll also see that not everyone is testing with the same "Problem size". So this is also effecting the outcome.
 
Re: **The Official TF LINPACK Thread**

If you'll look thru the results posted you'll also see that not everyone is testing with the same "Problem size". So this is also effecting the outcome.

I'm pretty sure problem size is dependent on available memory which is why most of us have different ones.
 
Status
Not open for further replies.
Back
Top Bottom