Intel 4.5GHz & 5GHz LinX Stable Club

[]siNNer[]

Nice job siNNer!

Thanks! I just wish i could find the issue in stabilizing past this speed, i keep getting bsod error 101 (vcore related) but it didnt get stable even at 1.475vcore. My temps are below 85ºC with that vcore but either i get that bsod error or my gflops drop to 55 from 67 expected at that speed.

[Ivanqu]

if you can explain how a 5ghz hexcore performs the same as my 4.1ghz quadcore!!!

? Picture?

[amrgb]

? Picture?

I don't have any around and I'm using the computer right now. But check owikh84's post: 980X@4.5Ghz gets 94.8Gflops.

http://www.xtremesystems.org/forums/...postcount=1163

Your 980X@5Ghz only gets 66.5Gflops, which is about what I get from mine at 4.3Ghz (not 4.1Ghz as I said before).

[Kimofil]

I don't have any around and I'm using the computer right now. But check owikh84's post: 980X@4.5Ghz gets 94.8Gflops.

http://www.xtremesystems.org/forums/...postcount=1163

Your 980X@5Ghz only gets 66.5Gflops, which is about what I get from mine at 4.3Ghz (not 4.1Ghz as I said before).

Man, HT is off on this mean more GFlops.
I think you know this.....

? Picture?

What version of Linpack library do you use fot the test?
Try download and run LinX it from 1st page of the topic-download included updated Linpack library by Intel whick fixed some GFlops issues.

[Leeghoofd]

Its water butn not chilled. GOLDEN chip anyway.

20°C on idle ? Must be chilled somehow (A/C unit) or pretty low ambients... Need to contact coolzero, coretemp's multiplier detection is off each time...

[Kimofil]

20°C on idle ? Must be chilled somehow (A/C unit) or pretty low ambients... Need to contact coolzero, coretemp's multiplier detection is off each time...

He posted in this thread that his ambient is +17C.

[donmarkoni]

Agree. That's as fake as they get. For so many obvious reasons that it's not even worth mentioning.

I don't know if I'd go as far as to say fake, probably something not quite right with the setup. Either way, with low GFlops the system isn't being near fully stressed.

I agree. This thread never had any accusations IIRC, so we should try to keep it that way.

If you can explain how a 5Ghz hexcore performs the same as my 4.1Ghz quadcore!!!

You answer is probably below:

Man, HT is off on this mean more GFlops.
I think you know this.....

[addressing to Ivanqu]
What version of Linpack library do you use fot the test?
Try download and run LinX it from 1st page of the topic-download included updated Linpack library by Intel whick fixed some GFlops issues.

It might be in conjunction with some other BIOS settings, which I'll try to investigate/replicate. Well, not at 5GHz.

20°C on idle ? Must be chilled somehow (A/C unit) or pretty low ambients... Need to contact coolzero, coretemp's multiplier detection is off each time...

He posted in this thread that his ambient is +17C.

My ambient was pretty low when posting this result: donmarkoni | i7 980X [6C 12T] @ 4522.6MHz | 1.4375 | Rampage III Extreme | Water | 3013A472
compared to this: donmarkoni | i7 980X [6C 6T] @ 4560MHz | 1.425 | Rampage III Extreme | Water | 3013A472
Joys of winter. I hope there will be more when temp drops below zero, if I don't die from pneumonia or something.

[Sniper]

Sniper | W3520 [4C 4T] @ 4851MHz | 1.488 | Gigabyte X58 Extreme | DDC, HK 3.0, PA120.3 + another 360
Sniper | W3520 [4C 8T] @ 4725MHz | 1.456 | Gigabyte X58 Extreme | DDC, HK 3.0, PA120.3 + another 360


Moved my rig into my sons room and opened the window tonight.
Ambient in the room was about 6 degrees, and probably lower for the air going into the rads.



Didnt have more time, so this was max for now with HT on:


2x360 rads.

[OC Nub]

^^Damn, thats a nice xeon you have. Great job.

[Sniper]

Thanks.

[SimpleTECH]

Sniper.... I am jealous...

Congrats.

[Ivanqu]

If i have time, I'll Try retest,

thanks all.

[CryptiK]

Sniper | W3520 [4C 4T] @ 4851MHz | 1.488 | Gigabyte X58 Extreme | DDC, HK 3.0, PA120.3 + another 360
Sniper | W3520 [4C 8T] @ 4725MHz | 1.456 | Gigabyte X58 Extreme | DDC, HK 3.0, PA120.3 + another 360


Moved my rig into my sons room and opened the window tonight.
Ambient in the room was about 6 degrees, and probably lower for the air going into the rads.

http://img.techpowerup.org/101204/linx4851.png

Didnt have more time, so this was max for now with HT on:

http://img.techpowerup.org/101204/linx4725ht.png

2x360 rads.

Really nice work there, I am envious of your boards max stable bclk. All these 220 - 230+ bclks, I can't go over 218 lol. One day I will have to get a better board and test mine in winter.

[Johnny87au]

@ sniper whats ur 24/7 overclock?

[Sam_oslo]

Sniper | W3520 [4C 4T] @ 4851MHz | 1.488 | Gigabyte X58 Extreme | DDC, HK 3.0, PA120.3 + another 360
Sniper | W3520 [4C 8T] @ 4725MHz | 1.456 | Gigabyte X58 Extreme | DDC, HK 3.0, PA120.3 + another 360


Moved my rig into my sons room and opened the window tonight.
Ambient in the room was about 6 degrees, and probably lower for the air going into the rads.

http://img.techpowerup.org/101204/linx4851.png[/IMG]

Didnt have more time, so this was max for now with HT on:
http://img.techpowerup.org/101204/linx4725ht.png[/IMG]

2x360 rads.

Good one

Long time no see, good to see the old man hasn't given up yet,

[donmarkoni]

If you can explain how a 5Ghz hexcore performs the same as my 4.1Ghz quadcore!!!

You answer is probably below:

Man, HT is off on this mean more GFlops.
I think you know this.....

[addressing to Ivanqu]
What version of Linpack library do you use fot the test?
Try download and run LinX it from 1st page of the topic-download included updated Linpack library by Intel whick fixed some GFlops issues.

It might be in conjunction with some other BIOS settings, which I'll try to investigate/replicate. Well, not at 5GHz.

Well, I tried to replicate low GFLOPS issue, but failed. No matter how I tried, didn't manage to get so low.
-running at 4700MHz HT OFF produced 101 GFLOPS
-disabling all CPU features dropped GFLOPS to 99
-using old LinX (linpack) dropped to 90 GFLOPS
-now with HT ON it should drop another 10-12 GFLOPS, didn't actually run, but comparing to my previous experience and observed results here.
-so, it comes down to 78-80 GFLOPS.
Only thing I know to kill GFLOPS totally is dropping RAM/UnCore to something like 1066/1600.

So, IDK


BTW, I did run full 20 loops at 4.7GHz

donmarkoni | i7 980X [6C 6T] @ 4703MHz | 1.475 | Rampage III Extreme | Water (ambient 9°C) | 3013A472

[acebmxer]

Well, I tried to replicate low GFLOPS issue, but failed. No matter how I tried, didn't manage to get so low.
-running at 4700MHz HT OFF produced 101 GFLOPS
-disabling all CPU features dropped GFLOPS to 99
-using old LinX (linpack) dropped to 90 GFLOPS
-now with HT ON it should drop another 10-12 GFLOPS, didn't actually run, but comparing to my previous experience and observed results here.
-so, it comes down to 78-80 GFLOPS.
Only thing I know to kill GFLOPS totally is dropping RAM/UnCore to something like 1066/1600.

So, IDK


BTW, I did run full 20 loops at 4.7GHz

donmarkoni | i7 980X [6C 6T] @ 4703MHz | 1.475 | Rampage III Extreme | Water (ambient 9°C) | 3013A472

i like that vcore voltage 33.00v?!?!?!

[Oztopher]

There should really be widespread recognition on how to run linpack properly. Like other people have said, i've seen so many 'stable' screenshots with poor gflops indicating that it's not been stressed thoroughly.

Everyone should read this - http://www.overclock.net/intel-cpus/...test-linx.html

Mind you, there are also a lot of respectable screenshots that i've seen with good amount of gflops on here.

Having varying gflops without awareness can also hinder your overclocking method, you might increase a voltage and it'd pass for longer, but it was only because of the low gflops. Therefore leading to a false idea of what makes it stable (or not).

[fordf250]

Does 4480mhz count? No? I dont think I like this club. LOL. Just kidding. 4.5G is a good number to use. I can get 21x213 stable and 195x23 stable but getting 1 more mhz from bclk at any multi to give the 4.5G is driving me insane. I am totally stable at 4480-4485mhz but 4500mhz causes instant lockup or bsod with any setting when linx starts.

The only nice thing about this is the bclk and cpu mhz max out at the same point.

[Gautam]

Well, I tried to replicate low GFLOPS issue, but failed. No matter how I tried, didn't manage to get so low.
-running at 4700MHz HT OFF produced 101 GFLOPS
-disabling all CPU features dropped GFLOPS to 99
-using old LinX (linpack) dropped to 90 GFLOPS
-now with HT ON it should drop another 10-12 GFLOPS, didn't actually run, but comparing to my previous experience and observed results here.
-so, it comes down to 78-80 GFLOPS.
Only thing I know to kill GFLOPS totally is dropping RAM/UnCore to something like 1066/1600.

It's because you're running with HT off. Check the other submission with HT on on Gulftowns in this thread, and you'll see that Ivanqu is well in line.

Newhit - 4725MHz - 44 GFLOPS
BobyTT - 4715MHz - 55 GFLOPS
Mydog - 4710MHz - 49 GFLOPS
myself - 4700MHz - 44 GFLOPS

Can't speak for the others, but FWIW I was running around 1800 for the memory, somewhere around 3GHz for uncore. Not terribly ambitious, but about as far as made sense to me to run 24/7. Not purposefully crippled, but nowhere near pushed as hard as possible. Chances are I was probably surfing the net at the same time as running it as well. Sure as hell wasn't making any attempt to break any GFLOPS records.

Now we could decide to be stability nazis and enforce shutting down all services and whatnot, but AFAIK, this thread is more about having fun than being some competition with heavy stipulations.

Of course, it doesn't surprise me that the bickering and calling people out as "cheaters" etc is occurring...it's bound to happen when people spend a lot of time and money to find a good chip and coax everything they can out of it, only to see someone else with a result that's better. Human nature, I guess, to try to dismiss it on any grounds possible.

[donmarkoni]

There should really be widespread recognition on how to run linpack properly. Like other people have said, i've seen so many 'stable' screenshots with poor gflops indicating that it's not been stressed thoroughly.

Everyone should read this - How to run LinPack stress test(LinX/IBT)properly-an explanation(maybe a guide)

Mind you, there are also a lot of respectable screenshots that i've seen with good amount of gflops on here.

Having varying gflops without awareness can also hinder your overclocking method, you might increase a voltage and it'd pass for longer, but it was only because of the low gflops. Therefore leading to a false idea of what makes it stable (or not).

I just started reading and I can say it's great article and a very detailed guide. Should be in first post. Bravo for finding/linking it.

Does 4480mhz count? No? I dont think I like this club. LOL. Just kidding. 4.5G is a good number to use. I can get 21x213 stable and 195x23 stable but getting 1 more mhz from bclk at any multi to give the 4.5G is driving me insane. I am totally stable at 4480-4485mhz but 4500mhz causes instant lockup or bsod with any setting when linx starts.

The only nice thing about this is the bclk and cpu mhz max out at the same point.

Well, that's a nice overclock, nevertheless. PM me your BIOS template, or point me to thread/post with it, so I can take a look. Four eyes see better then two.

[donmarkoni]

It's because you're running with HT off. Check the other submission with HT on on Gulftowns in this thread, and you'll see that Ivanqu is well in line.

Newhit - 4725MHz - 44 GFLOPS
BobyTT - 4715MHz - 55 GFLOPS
Mydog - 4710MHz - 49 GFLOPS
myself - 4700MHz - 44 GFLOPS

Can't speak for the others, but FWIW I was running around 1800 for the memory, somewhere around 3GHz for uncore. Not terribly ambitious, but about as far as made sense to me to run 24/7. Not purposefully crippled, but nowhere near pushed as hard as possible. Chances are I was probably surfing the net at the same time as running it as well. Sure as hell wasn't making any attempt to break any GFLOPS records.

Now we could decide to be stability nazis and enforce shutting down all services and whatnot, but AFAIK, this thread is more about having fun than being some competition with heavy stipulations.

Of course, it doesn't surprise me that the bickering and calling people out as "cheaters" etc is occurring...it's bound to happen when people spend a lot of time and money to find a good chip and coax everything they can out of it, only to see someone else with a result that's better. Human nature, I guess, to try to dismiss it on any grounds possible.

I know, but with HT ON I got 87 GFLOPS at 4522MHz, much more then 40-some to 50-some.
I understand what you talking, and that you are talking on your behalf, but as someone said, people are looking here for good batch# and try to find ones for themselves believing in fake stability.
Don't misunderstand me, I'm not calling anyone a cheater, and have even reacted against such accusations, but IMO it's not a real stability with such a low GFLOPS. Don't get me wrong, I'm not underestimating anyone's efforts.
I wouldn't even want anything changed here, cause this i just a club, no disrespect meant, on the contrary, we should have fun here and club should serve as a guide.
That article, How to run LinPack stress test(LinX/IBT)properly-an explanation(maybe a guide), should be used for guideline, but I wouldn't go that far on disabling win services, AV, or similar, but people should avoid disk activity at any cost and just restrain oneself from using computer for less then an hour.
I am not satisfied if knowing that my system is potentially unstable and that I could avoid it by letting it be properly stressed, cause I fold on it. Just a side note: I'm not a stressing freak, this is the most useless stressing I've ever done, 20 loops of LinX.

EDIT: Sorry for double posting.

[Oztopher]

I agree.

Quite a few people have got 40-50 glfops on their 980X at 4.5ghz +, with HT enabled. Yep, quite a few people, that must make it right.

However, look at runs like this - http://www.xtremesystems.org/forums/...9&postcount=40
and this - http://www.xtremesystems.org/forums/...&postcount=423

Or even this - http://www.xtremesystems.org/forums/...&postcount=700

He's even using an older version of linx. So you can tell there's a lot of variations in gflops between systems, or even between different boot ups Guides like i linked can make sure you're running it consistently everytime, and hopefully even between different systems.

I wouldn't say it was cheating either, just a lack of awareness of how to make sure linpack is properly stressing the CPU. This is why i sometimes use prime95, as it's consistent. But of course, if you know how to get linpack running consistent then it really can be a hell of a stress tester, which i've realised recently

I don't disable any services either, that's a bit over the top and you don't really get back your efforts

[some_one]

It's because you're running with HT off.

Ummm, maybe donmarkoni would be kind enough to do 3 runs @ 4GHz of 64-bit with say 4GB of mem (24000). I would expect ~90% of full SSE with no HT (6 threads) and ~75% of full SSE with HT (12 threads).

HT = ~72GFlops
No HT = ~86GFlops

FWIW even though using SMT (HT) gives lower Flops it uses more power, ~10% more on my 860, so is probably more stressful than the higher Flops on separate physical cores.

[fordf250]

I have even made it run wrong with an i7 950. Twice now trying to get 4.5G Linx has put out a 40 gflops pass showing 100% load and 75C max temp on air at 1.525Vcore. Stopping linx and restarting it sends it to the 64-65 gflops 100% load and 92C it normally runs. I didnt read how to run linx the right way yet but its not hard to tell 40gflops, 75C and 1.525Vcore on air is wrong.