ran mine with HT off and did 100Gflops, with un-modded 6.4 Linx. maybe 6.5 will integrate the new binaries.
Printable View
ran mine with HT off and did 100Gflops, with un-modded 6.4 Linx. maybe 6.5 will integrate the new binaries.
I use to use Orthos for 24hrs for stability testing.
So how long or how many passes are needed for LinX for the same level of stability
TIA
Zeus
A question that can never be answered without a lot of argument :p:Quote:
Originally Posted by Zeus...
PERSONALLY (key word there, it is in all caps in case anyone didn't see) I have never had any stability issues with an OC that can run 20 passes of LinX at "ALL" memory usage. I run 20 passes (takes about 1 hour give or take) and then F@H has never, ever had a crash or error.
Cheers m8 I fully understand the stability issue but a rough idea of what users find stable is a good startQuote:
Originally Posted by tool_462
This "test hours" vs "stability" is like talking about the sex of the angels. :D
The theorys of ..... "oh I test 12h because 1h is not enougth" .... "oh I test 24h because 12h is not enougth" ..... I understand what people are saying .... but by the those same principles I ask ... and if @ 24h and 1 minute the test fails ? :rofl:
What I've learned in some years of OCing my PC's is that the only real test is the dailly use that will come ahead. For me the tests I make is just to find a level where my PC is ready to begin to face the dailly base work load.
If in the process of the dailly use something feels not being stable, then I'll give it a toutch :up:
Quote:
Originally Posted by jVIDIA
Because if 24 hr passes, your chances of 24:01 failing are slim to none.
Your chances of 1:01 failing after 1hr passing is totally higher.
The idea of 24 hr passing means you can guarantee 24 hr stability. What do you guarantee after 1 hr stability? 1 game of starcraft 2 before your game crashes? :D
On a separate note, does GFlops change a lot? I got like 36 GFlops only last night testing my i930 at 3.8ghz. Not sure why that is. It was doing 48 earlier. I'm concerned because my temps were hovering int he 60s only. Low 60s. Made me think my CPU wasn't working as hard as it should be able to!
Edit: I see it a few pages back. Makes me wonder! Maybe I should install Prime 95 just in case! Or maybe OCCT's Linpack test is decent.
Edit 2: It seems OCCT Linpack pushes my temps up further. I rebooted my system and ran Linx and after 30 minutes I recorded the max temps from RealTemp. Now running OCCT and my max temps have gone up a degree across the board.
In the good old days, memtest was the bread and butter of stability testing. What I really liked about it was that you booted it from disk, and then, no matter what it would not corrupt your OS.
Is there a way of getting a bootable linpack?
For some reason on my Xeon X3380 every time i hit ALL mem gives me an error of NOT ENOUGH MEM...this happened to me on the UD3P and now on the USB3P...i have to use 19k problem size and goes fine, with my Q9650 this didnt happen...any clues ? same with IBT...
IM using WIN7 64bits
Do you have page file disabled? If yes then thats your problem.Quote:
Originally Posted by andressergio
thanks bud i do have it disabled...will re tryQuote:
Originally Posted by ChaosAD
Cheers :up:
Sergio
Glad i could help, running linx without page file is no issue, you just have to run a smaller problem size. With 6gb ram running 25000 problem size is good enough :up:Quote:
Originally Posted by andressergio
It would be nice if there could be some sort of reproducibility of a failing testcase. Chances are the specific data pattern that triggered the fail would repeatedly fail with voltage and temperature being constant. If we could dump the testcase, it would greatly reduce testing time of finding a semi-stable overclock. Obviously continued operation is still necessary to obtain maximum coverage still.
hmm some question
i have a 930i7 at 4.0ghz for daily usage but i am having a quite amount of heat in here cuz my ambient is not that good..around 35+ so my settings for 4.0ghz seems to stable at linx 768mb mem but only 10 loops...but it fails on 1536mb mem on 10 loop....also the temp of the 768mb full load is around 89c and 91c for 1536mb...i am goin to fix this damn temp by getting better fans for my casing but my question is, does stable in 768mb 10 loops can give stability on daily usage? i mean is it enuf? or is it the temp that giving me problem to get stable in 1536mb linx?
No.Quote:
Originally Posted by intothefantasy
It's likely.Quote:
Originally Posted by intothefantasy
Even if it's stable in LinX for 24 hours, doesn't guarantee that your computer will be completely stable for 24/7 use, but typically 20-50 runs guarantees about 80% stability.Quote:
Originally Posted by intothefantasy
I'm having problems with the program. Recently I formated the computer and since then, it did not work anymore. Someone could tell me the reason for not working? At one point, the program came to run smoothly, but was not stressing the CPU. The program ran smoothly, but not stressed.
I tried to download version 0.6.4 and 0.6.3, both versions with the same problem. This problem is occurring in the program OCCT (Linpack) also.
http://img27.imageshack.us/img27/4065/linxerro.png
@ EDIT:
It seems that the problem was solved, because the moment is running smoothly, as pictured above. But is not stressing the CPU.
Does anyone know why?
http://img9.imageshack.us/img9/2631/linx3.png http://img251.imageshack.us/img251/7812/linxerro2.png
Do you have page file enabled?
Run the program as administratorQuote:
Originally Posted by _AntLionBR_
create a shortcut for the program, right click the shortcut and select properties. Under the compatibility tab in the properties window select the run as administrator check box and then apply the changes..
Quote:
Originally Posted by zalbard
I made these procedures, and no success.Quote:
Originally Posted by Major
I decided to uninstall the antivirus program (Comodo) from the system and feels like the problem is resolved. The moment is stressing the CPU with no problems. Thank you all! :up:
Ack! So Prime95 for 24h is still the way to go for 100% stability testing? Jeez I wish there was a quicker way.
i know if it passes linx for 20 mins its fine, if i have a crash i can tell what it is and adjust accordingly, linx from my experience isnt the best for testing IMC, ive had runs pass linx np's but crash during heavy demanding online gaming (bc2, mechwarrior living legends), in which case more vtt or ram timing adjustment fixes IMC stability..
i dont see the point of stressing my cpus that much for that long (hours) when i can get my system stable with far less time and stress to my chip
True, true. I'm running 20 passes of LinX and calling it a night. If it passes I'll do a bit more tweaking tomorrow to maybe get the volts down, if it doesn't pass, I'll still do more tweaking.
I hate not being able to use my desktop for long hours while Priming. I think I'm gonna go the Linx 1 or 2 hour test and then some gaming (BC2). The only thing is the whole corrupting Windows while OC'ing thing. I prefer to OC before getting my Windows install fully restored in case things go haywire.
Plus, whirring fans at night and a case lit like the sun (note to self: dont buy led case fans) its annoying to sleep while stress testing.
thanks for this app, prime does put me off since having to use it for long lengths of time.
running this now set to 20 passes and all my ram, although all my ram is only 2.9gig out of 4 gig, it wont test ram in use by the OS?
seems fast enough already 2 passes in 5 minutes.
Just downloaded it, ran it and passed all 20 tests , says my glops is 51, 51, 51.. Doesnt change at all... Stable right??? Also with the problem size do i just leave it default?
The problem size depends on the amount of memory you want to use. Selecting ALL helps get the highest amount possible.Quote:
Originally Posted by Johnny87au
Usually you want to aim for ~1GB less than your total memory (ex: 6GB (5GB used), 4GB (3GB used)). The more the better.
ok thanks dude, Ill stick to clicking ALL
mine passed all 20 so I hope my stability problems on this board are behind me now, I had 40 gflops rating.
Sorry for being ignorant but could somebody explain "GFlops" to me.
Do i want it to be higher or lower?
Thanks in advance.
higher is better, though difference betweeneg 45 and 53 is not reflected in eg benchmarks like 3D06, Pi32, PCmark05 etc... on my I7 setup the Gflop rating is very dependant on the CPU features setting
can this burn a laptop???
or in any way actually damage it?
Any stress test can cause damage 'if' the cooling isn't good enough and the components overheat. Doesn't matter if its a lappy, desktop or net-book.
But if it cant handle it is it any good? Plus some do have a throttle that decreases the voltage multipliers when things start getting too hot.
Could you explain what changes has been made to linpack_xeon64.exe and linpack_xeon32.exe? The digital signatures from Intel has become obsolete. I don't suppose it's some clean actions?
There is a newer version of the Linpack Library on Intel's website, that's what I've been using for a while now rather than the one included in LinX 0.6.4.
Here are the Digital Signatures on both xeon64.exe and xeon32.exe with the newest Linpack.
http://i26.tinypic.com/13z0jtu.jpg
http://i27.tinypic.com/2mhzm9x.jpg
For those running the i7 980X with HT on, using the latest Linpack should result in higher and more sustained GFlops.
You can download the newest Linpack Library from here:
http://software.intel.com/en-us/arti...pack-download/
Simply just copy xeon64.exe and xeon32.exe and replace the originals in your LinX/IntelBurn directories.
Great tip, thanks. :up:
Maybe the 4.5G club needs a clean-up.Quote:
Originally Posted by zalbard
There are some bogus results in it, esp. the 6C12T ones.
For intel linpack, low Flops=less Stress, and the difference could be very big.
It is possible to pass linpack by preventing it from running efficient.
Please update Linx with the newer linpack files. If one uses LinX 6.4 without changing the Linpack files to the one listed above you get insane Gflops with Gulftown without having to fully load your cpu.
Old VS new Linpack files
When I run LinX after about 5 mins my videocard drivers stop responding and the screen flashes black continuously. What kind of sign would this be?
Win 7 64bit
759 classified
i7 920
Try bumping VTT one notch higher.Quote:
Originally Posted by Big Myke
Correct me if I'm wrong but isn't it ideal to run linx .64 in safemode to get the MOST memory stressed during the test? The only problem I expierence while testing in safemode is that realtemp or other temperature apps do not work in safemode. :shrug:
After booting into W7x64 I have 5GB available to stress test. IMO, it's enough.Quote:
Originally Posted by Big Myke
If I start LinX and then stop it after 20 seconds or so, I have 5.5GB available, as Win releases some memory.
And that downside is a big one.
If you want to be able to stress more memory disable all the aero related stuff and extra services you don't need. If you don't want to do that install windows xp x64 it uses very little memory.Quote:
Originally Posted by Big Myke
I notice with using the new Intel versions of Linpack my dual 5630 Xeons are stressed more. Highly recommend a new version of this tool with those new compiles from Intel in it.
harder to pass new library
I found that running only the # of threads matching your physical cores actually puts more strain on Gulftown/Westmere-EP CPUs. For example, running 12 Threads instead of 24 on my Dual X5680 gives higher Gflop rating and produces higher temperatures than running with 24 Threads.
Apparently WIndows 7 is intelligent enough to put the load only on the physical cores if the # of threads don't exceed them.
Next I will follow Mitchell's great tip and update the Linpack librariers. Curious to see the difference :)
Edit: Nope, updating Linpack libraries is useless. I get next to no CPU load and 50 (!) Gflops rating :ROTF:
Running 12 threads is still the best option for dual socket 1366 and linpack.
@jcool thanks for the heads up m8! :)
No problem mate, but see edit.. updating libraries does not solve the HT problem. Just run 12 threads... or prime x64 with all ram, which is VERY sensitive as well.
Edit: New libraries are more sensitive though. I just got an error with the same settings that passed on the old libraries :)
About gflops plz see http://www.xtremesystems.org/forums/...d.php?t=256633
Yes, we've known about that for a long time. But seeing as I normally run with HT, since WCG gets a 40% boost from it, I don't want to disable it just for stress-testing. Hence my tip that you can simply run LinX with the # of threads corresponding to the amount of physical cores in your system, and you get the same effect as with turning HT off.Quote:
Originally Posted by sniper_sung
Here's a shot with updated libraries, but otherwise identical settings to my earlier screenshot - Gflops are lower, but the program is even more sensitive now. Looks like I have some fine-tuning left to do ;)
Afaik HT lowers per core performace but 'raises' the core count. Meaning a i7 with HT off will have performace rating of 100 on 1 core meaning 400 on four and with HT on Per Core PFRating drops to 75 but core count jups to 8. Meaning Anything that uses up to 4 Cores will be faster on a non HT's i7 but anything that uses more than 4 cores will benefit from it.Quote:
Originally Posted by sniper_sung
Bold explanation.
An 4 Core Usage aplication =
4 Non HT'd cores 400
4 HT'd Cores 300
An 6 Core Usage aplication =
4 Non HT'd Cores 400
6 HT'd cores 450
An 8 Core Usage Aplication =
4 Non HT'd Cores 400
8 HT'd cores 600
This is of course only in case the Aplication Code is superb and it really uses the cores.
Simply taken. Video Compression, WCG , Rendering ... HT will profit
Anything Else, HT will drop performace.
That's why we see a Gflops Drop in Linx with HT on.
I observed 4 GHz i7 930 with HT on gives me about 51 ~ 52 GFLOPS , HT off gives me around 58 ~ 60
I talk too much
Hey guys I have a bit of a dilemma here. Currently I have my Q9550 running at a stable 3.825ghz (450x8.5) I can get into windows at 4ghz (471x8.5) but when I run LinX after the first pass.. cores 0 and 1 would drop 8-10c even though they are still under 100% load. Cores 2 and 3 would soon follow after a minute or so. Gflops would drop and times would sky rocket. Temps are around 77 77 72 72. Some say I am throttling but I have seen people run these temps under LinX without issues. However I did pass OCCT 1 hour without errors at 4ghz and temps never touched 70c. Am I at the limit of my board? What else can I do? Any ideas?
heres a screenshot
Real Temp shows only 78% load but the task manager and core temp show 100% load on all cores.
safan80
thanks for the info mate
@orlywutlol
Maybe your CPU is throttling and you don't know it. ;)
RealTemp is proven to be the most accurate, so if it says your CPU usage is falling, you bet it is.
If you could post your system specs and BIOS template, it would be helpful, but you probably should put in your board's thread, or better start a new thread, and not go of topic here.
Does this program use single or double precision?Quote:
Originally Posted by Dua|ist
Is there a way to disable SSE?
Is there a way to disable Intel MKL?
Is there a way to test it using AMD Core Math Libraries (ACML)?
Does this use OpenMP or MPI for parallel processing?
Can you give the option of running this just once (instead of a minimum 3 times)?
This is easier to use than the actual LINPACK/HPL though.
I managed to get LINPACK to compile using g77 ok, but it was saying that I was only getting like 460 (double precision) MFLOPS per core on an AMD Opteron 2220 SE (2.8 GHz, dual-core).
I haven't tried any of the other compilers yet, and I haven't tried enabling SSE.
I also haven't tried linking it with Intel MKL or AMD AMCL yet either.
HPL uses MPI, so I'll have to test that out too. (Especially with HP MPI, Intel MPI, MPICH2, and OpenMPI).
At this point you would be better using IBT 2.5
LinX hasnt been updated in a long time
Windows task manager shows a scheduling load not a CPU load.Quote:
Originally Posted by orlywutlol
why do you say that? You can just download the latest linpack files from intel and replace them. LinX is just a gui for those files anyways and it works great.Quote:
Originally Posted by slim142
Really?
I didnt know about that. Good to know that you can keep the libraries up to date.
Now, IBT is not that bad either. The UI is great too. Just wished Dualist was still here to keep on improving the program.
+1Quote:
Originally Posted by safan80
Just watch for library updates.
link and stuff? :)Quote:
Originally Posted by safan80
try page 30 the link is there.Quote:
Originally Posted by dMITIj
Check the OP in the thread in my sig. I uploaded LinX with updated Linpack files a while ago. :up:Quote:
Originally Posted by dMITIj
hi! sorry for the change of subject... i was wondering if there was any way to force a determined amount of memory usage in LinX.
i mean, just like in Prime95, where you just set the amount required and the program automatically frees up memory for you..
thanks :up:
when I run Linx is set in the configuration that was a high priority / or load (can not remember what exactly). Set in the necessary things about the "problem of size" and "memory". At the end push the "start" button, the timer turns on and kind of testing program. CPU usage does not even exist and I don't have this:
http://img233.imageshack.us/img233/6...2009192622.png
My OS: Windows 7 64 bit.
P.S. I use administrator privileges. Sorry for my english :]
Try using ALL memory button. It is so short problem size you can't even see the load on CPU for the fraction of a second.
I had posted this question in the forum for the Asus Rampage III Extreme, which is the Motherboard I am using, but this seems like the correct place to ask, so I'll post here with some more data.
Having issues with my new Build, i7 980X, Rampage III Extreme, 12Gb of Gkill RAM. LinX uses 9280megs.
My problem is, with overclocking and stabilizing the system.
I've tried using the following: Prime95, LinX and Cinbench 11.5
Cinebench works always, (its quick)
Prime95 runs for hours (longest run was 9.5 hours)
Runs from 3333 all the way to 4300 without errors, after tuning
LinX, works something and yet not others, THIS IS THE PROBLEM
On some boot it runs at 81+Gflops, other times 51Gflops, this is regardless of the CPU/Memory speed, in fact at:
4200 with Memory at 1728MHZ, it does 81 on some reboots, when it does this LinX will run without issues for 20 runs, if after a reboot it is 51Gflops, then it normally dies with in 3 runs. This without change options only rebooting.
In both cases Prime95 runs without issues (for over an hour). Cinebench shows no change in performance when I run it no matter the speed of LinX. Nor does SiSoftware Sandra show any change in performance.
When it runs at 81Gflops temps may reach as high as 75c, when it runs at 51Gflops, they never exceed 70c, so even though it is at 100% CPU, there does seem to be less work being done.
So I'm confused if it is a LinX issue, a MB issue or what as I am at a loss to understand why LinX only works sometimes and then returns far better results then when it does not work. This is making stability impossible with LinX since having it fail means nothing if something else is wrong, which affects performance of its results by 30Gflops, but only for LinX.
ERIC
Is Dua|ist going to update this program anymore? It does get old asking if users are using the latest linpank with it.
Thanks the updated version helped fix the random changes.
ERIC
you trusty tech site has a mirror
http://techpowerup.com/downloads/1902/file.html
It's just a UI, so doesn't matter too much for the most part. Just keep updating the binaries like I do. :up:Quote:
Originally Posted by safan80
Thanks :up:Quote:
Originally Posted by zalbard
I have a problem with LinX, if I click on the button "All" to memory, I get an error "Stopped Unexpectedly (not enougth memory ?)...
Why?Quote:
Intel(R) LINPACK 64-bit data - LinX 0.6.4
Current date/time: Thu Nov 18 15:48:12 2010
CPU frequency: 3.038 GHz
Number of CPUs: 12
Number of threads: 12
Parameters are set to:
Number of tests : 1
Number of equations to solve (problem size) : 26832
Leading dimension of array : 26840
Number of trials to run : 10
Data alignment value (in Kbytes) : 4
No runs is allowed for C:\Divers\utilitaire\bench\linx_0.6.4\linpack_xeon 64.exe: not enough memory
and is it possible to have AIDA64 (replacing Everest) for monitoring?
thanks ;)
Your missing the point. I know what binaries I use, but I don't know what binaries others are using and it gets old asking.Quote:
Originally Posted by zalbard
Hi guys.
As IntelBurnTest also seems to use Linpack, is it fully possible to update the Linpack library files in IntelBurnTest with the latest Intel ones the same way you can with LinX?
Thanks for responding.
Sylver123
It is. :)
IntelBurnTest is version 2.5, you can update the librairy of LinX 0.64 or it's useless?
thanks ;)
Should be interesting when Sandy comes. My guess is that an SB quad at 4GHz clock and using 1024MB of RAM with Linpack will hit 100GFlops. :D
You need at least 100MB virtual memory !Quote:
Originally Posted by tistou77
http://www.imgplace.com/img521/8285/36naamloos.th.jpg
hey guys i need a tip: i tried many settings in my BIOS but every time i got low GFLops ... i never passed 50 ... so i need tips ... TIA.
it takes at least 100mb of virtual memory that operates LinX?Quote:
Originally Posted by AceBaran
I turned off my virtual memory...
there was no way around this?
thanks ;)
You need at least a small swap file to use the max memory button, yes. However, you can just select a slightly smaller amount of RAM, and it will work just fine as is.Quote:
Originally Posted by tistou77
ok thanks ;)Quote:
Originally Posted by zalbard
Well i might have dramatized it a bit.. maybe not 100mbQuote:
Originally Posted by tistou77
Like Zalbard says you need at least some virtual ramm..
Great program. :up:
Having troubles getting 0.6.4 to work with my 980X rig. Works fine with the Sandy, I searched this thread and found the the required Linpack updates but it refuses to start stressing. Tried version 0.6.3 here and it worked. Anyone have any ideas?
TIA
Make sure the user who started linx.exe has the privilege to write into the folder.Quote:
Originally Posted by crossg
Is anyone able to contact Dualist?
or check the russian forum where he is at? Maybe he started working on this again? or made a recent announcement?
I really liked how this program was progressing
He hasn't been on-line for around 6 months there.Quote:
Originally Posted by slim142
Sucks to hear! :(Quote:
Originally Posted by zalbard
Thanks.:) Checked the user privileges and they look good, same as version 0.6.3. I will use 0.63 with the updated linpacks for now.Quote:
Originally Posted by sniper_sung
Latest LinPack Updated to w_lpk_p_10.3.2.005 (2/3/2011)
Massive Gflops for P67 now!!
Pass it around and remind everyone of the new update, yes this is with SP1 (Win7)
http://i.imgur.com/cSH6h.png
http://software.intel.com/en-us/arti...pack-download/
old news.. I've ran this 2 weeks back
5.0GHz LinX 8 Threads @ 84GFlops
http://i458.photobucket.com/albums/q...5084GFlops.jpg
5.0GHz LinX 4 Threads @ 132GFlops
http://i458.photobucket.com/albums/q...HToffAIoff.jpg
A bit off-topic, but what Windows skin is that? I like it :up:Quote:
Originally Posted by Lsdmeasap
Guess I missed it :DQuote:
Originally Posted by owikh84
I did see people mentioning to update again, but thought they meant to update to 10.3.0.003
I'll let you know once I am back in that system, I'm working on reviews right now so I'm not in that setup.Quote:
Originally Posted by danielkza
I believe it is this one but I do modify my own, so it's likely a mix of a few
http://benjamin905.deviantart.com/ar...win7-151195376
so we need to update the 2 files into the linx folder? anyway i've just testd it ... with the older version i got many gflops that the newer ... (just testing my system @ default)Quote:
Originally Posted by Lsdmeasap
Quote:
Originally Posted by Lsdmeasap
So how do you use this???
Just get the most recent version from the OP in my LinX thread, everything's already integrated there.
Hello to all.
I'm try to run Linx to my PC with 8Gb of RAM, all others suite works fine, but Linx crash after few seconds with this error (if I set more than 4100Mb of memory)
http://www.webalice.it/andyzanny/linx_error.jpg
So, thinking there was something wrong on my configuration, I tried also tu run it on the other PC always with win 7 x64 and 8Gb of RAM, same result
I tried to activate the swap file but no help
Any thought?
Tks
is your system memory test stable?
New Intel Linpack, 10.3.5.008.
LinX updated: linx(0.64)_10.3.5.008.zip
So for a starter, of LINX stability tests, can someone explain me what should i do with the Linpack? from the first link : / ?
Or i should just use the LINX update file from the second link? thanks guys.