PCMark Vantage World Records

[SteveRo]

The 32nm dual cores hitting 7GHz wouldn't do much for this benchmark as the would probably need to be around 6.3-6.5GHz stable to run the bench. Where as I can just run my 980x at 5.9GHz+ and easily beat the dual core at 6.5GHz. As far as what Steve will hit on a single stage really depends on the chip some chips scale some don't. I had a few chips on single stage that only ran 5.3-5.4.

I will be happy if I can do pcmv at 5.3-5.4
I think that might add a little bit to my score maybe

[GullLars]

i've heard electronic cicuit runs better on cold temperature, and see now why this 'kid' from the sub-zero cooling section here at the forum is building a 3.5-4.5KW cooling solution,
if a 920 running 4.5 can draw ~300-350? W then i'm not sure if a 980 (stock 3.33) would be able to reach 5 so easily with a 300W cooling mechanism,
taking u'r chip into account, a 3.46Ghz which has to have -101 to reach 5.6, i feel like a fool saying what i did few posts back

Could this "kid" be making a (cascade?) cooler/compressor setup to make his own L2N for running 24/7?
Such a 3,5-4,5KW thing could be nice in the winter if you live north of the polar circle :P

One of the reasons CPUs scale with low sub-zero temperatures is the effect of temperature on resistance in conductors. Another may be reduced leakage, tunnelling, or something of that sort. (i haven't given the second a thought yet)

[Computurd]

superconductivity.

[mike]

i think actually, these guys over there has already gave they're best effort with they're setups, now what they'll need to do, is probably arrange another devices, RC and/or disks..
you seem to be a bit skeptical of what u have in your hands, yet, don't forget these results are even before running it on an SS unit .

You think I am tapped out already? ( assuming you were referring to me as well?)

Only one way to find out....

Excellent work Mike, congrats.
Wow I think I got to keep the WR all of 48 hrs this time!

correction - actually 54 hrs!

Thanks - grats to you as well for keeping me on my toes... Hate to tell ya - there is at least 6k worth of points left within reach for me

[GullLars]

@Computurd: For that you would be talking L2N or Helium... You need to get into double digit Kelvin, prefferably low double digits, to get superconductivity. Most materials need single digit kelvin or very low double to become supercondictive, so you'd need helium (boils at ca 4 Kelvin).

edit:
Mike, it would be awesome to see you push past 40K points whitout "cheating" with system RAM cache exploits and that sort of thing. Keep it up, and go embarress Corsair Labs for their PR stunt by trying to beat you

[SteveRo]

Hate to tell ya - there is at least 6k worth of points left within reach for me

Don't tell me that!
You need to keep me guessing!

[GullLars]

Don't tell me that!
You need to keep me guessing!

Hahaha, it's like SPOILER!

It's really fun to watch you guys go at it with ridiculus setups, Steve with almost full disclosure of all settings and hardware, and Mike just lurking in the shadows and no more hardware details than is listed at ORB.

[SteveRo]

Yes, "almost full disclosure" is right.
Like Mike, I also keep a secret or two.
But probably not 6k worth!

[onex]

You think I am tapped out already? ( assuming you were referring to me as well?)

Only one way to find out....

lol man, you definitely hold some secrets
i was thinking u we're done .

Could this "kid" be making a (cascade?) cooler/compressor setup to make his own L2N for running 24/7?
Such a 3,5-4,5KW thing could be nice in the winter if you live north of the polar circle :P

i don't know yet, with SUCH system, it well can be,
he says he lives in Norway btw, might be setting u a contact for good prices , Anvil was asking about it few pages back here.
i think a system should be ~600$ for a -40 such as Sdumper one (and up going).

One of the reasons CPUs scale with low sub-zero temperatures is the effect of temperature on resistance in conductors. Another may be reduced leakage, tunnelling, or something of that sort. (i haven't given the second a thought yet)

i'm thinking heat radiation or different affects on an electric pulse messing around processor potential productivity,

superconductivity

we'll have to go further into it .

[lowfat]

Yes, "almost full disclosure" is right.
Like Mike, I also keep a secret or two.
But probably not 6k worth!

I think it is time for you to find some LN2 if you plan on trying to keep this record

[eva2000]

wow 35k mike crazy

i'm further down list still with just single 80GB X25-M 23,308 http://service.futuremark.com/compare?pcmv=309479

why is my cpu = missing cpu labeled ???

[onex]

I think it is time for you to find some LN2 if you plan on trying to keep this record

no.. it's not about that,
Mike (for revealing his secret) O.K, NO.. EDITED
he's not using even half of the HW others are using.., (at least by what he said).
it's about tweaking it further and further,
no clue what the guy is doing there, yet surely, it seems he KNOWS what he's doing to be reaching such "grades"..

as for that "kiddo", i just recalled, he was asking for some cooling refrigerant, he's not using LN2.
on an SS, with massive compressor and condenser afai can tell, they were suggesting him iirc 404B.

[mike]

wow 35k mike crazy

i'm further down list still with just single 80GB X25-M 23,308 http://service.futuremark.com/compare?pcmv=309479

why is my cpu = missing cpu labeled ???

I am a bit confused about that as well. I have a couple of theories for that.

My first theory in my case is that the 980x that I am using is a third dead. The first channel of my memory controler is dead, and even the last 2 channel that still do work can only run to up to 1800mhz max for memory speeds. ( hence my last runs were done on only 4GB of memory)

The second theory is that because I am using an early version of Windows, some file could be missing, that FM system needs to properly detect the CPU. Since I haven't updated WIndows, whatever is missing, I have a missing CPU.

I had attempted to switch between various Processors on that install, Retail 980X, ES 980X and even retail 920 bloomfield. No matter what FM would see it as "MISSING CPU". Can't even detect 920....

Utterly confused about that.

[mike]

edit:
Mike, it would be awesome to see you push past 40K points whitout "cheating" with system RAM cache exploits and that sort of thing. Keep it up, and go embarress Corsair Labs for their PR stunt by trying to beat you

No worries about me using system RAM cache exploits. I have been an opponent for using such a tweak for a long time.

RE: 40k - This will of course involve the use of LN2, and a slightly different setup. My Areca 1231ML is also limping on it's last leg. I had killed that controler twice trying to mess with it to somehow get it to perfrom better. AT this point after killing 8 out of my 12 SATA channels it has reached the end of it's road. 2 Acards won't be good enough anymore if I want to remain competitive for the top spot . I want to explore using SSDs soon, seems like there might be a chance to give the Acard ANS9010 a run for the money with SSDs when properly setup.
Sucks that borked my Areca, would have loved to have a bit of $$ from the resale of it to help me fund my next controler....

Stevero is getting much too fast now, especially with his SS coming in soon.....

[GullLars]

Could it be the mobo?

[mike]

Don't tell me that!
You need to keep me guessing!

Sorry for the Spoiler Steve. I had let someone else know who wants to run for the top spot again, that 40k is the minimum requirement to beat me, thought I'd let you know too.

[mike]

Could it be the mobo?

Hmm, maybe but doubt it. My CPU however is bad on various mobo, so the missing memory channel is for sure a CPU related problem.

[onex]

steve,
does u'r ML has any cache backup in any chance of a sudden electrical failure?

i was wondering over this as 4GB of data is quite A LOT to miss..

[mike]

steve,
does u'r ML has any cache backup in any chance of a sudden electrical failure?

i was wondering over this as 4GB of data is quite A LOT to miss..

Not to talk for Steve, but in my case after using the 1231ML for several years, I have never lost any data due to the cache.

However I also do take precautions. When capturing HD video I usually turn the chaching off, to prevent any costly loss of data. ( but honestly most of the time I forget - and still never had problems...)

However when editing video, or graphics I always switch the caching back on. Editing video, using programs like Adobe After Effects, Photoshop etc. benefit greatly from this!




PCMark Vantage simulates Graphics editing in test1, and video editing in Test 2. Hardware cache is good precisely for that. Software caching very bad for that in the real world.
Too bad PCMark Vantage does not credit system memory better in it's point system. My biggest complaint with this benchmark is precisely the lack of fully utilizing all system memory.

[SteveRo]

Good afternoon Guys,

I really haven't had the chance to lose any data due to cache because I use all my spare time benching - trying to catch Mike!

[mike]

First thing to do is see how much it scaled with the better cooling.

Use same volts you are running currently on your chiller and slowly increase FSB. Run PCMV CPU tests and see until when it will still pass.

With that you gonna have a good picture on how is the scaling on your cpu.

After that I would start increasing vcore and a tad of your vtt, pll can stay in that area 1.9 to 2.

You will reach a limit there when the SS won't be sacling anymore with volts because the heat will peak quickly during PCMV cpu tests, and thats when you will notice that none of the volts will help you with fsb.

PS.: You have a UD7 did you use the mod hicookie suggested?

Excellent advice! My only recommendation would be to actually lower the PLL slighty- rather to 1.7 or maybe even sligthly less.


STEVERO - please make sure to put a probe on your CPU as well and try to monitor the IHS temp as well. The evap could be -39C, yet the cpu itself could be in the positives. 1.55V should be a nice single stage vcore to start with. 1.7V sounds potentially high.

[SteveRo]

@ Mike - thanks, will do, question - where do you like to mount your probe - just beside the heat spreader?
Also, question - cpupll intel "absolute max" is 2v, why drop to 1.7 when 1.9 is < 2v - lower temps?

[Computurd]

PLL can cause instability on the 1366 procs..most recommend lowering it to achieve stability...
i hate to link this guide, and wish i knew of a better one as i hate this forum, but anyway....

VOLTAGE GUIDE

anyone with a better guide please sound off!

[mbreslin]

Interesting guide thanks ct, on a related note I fail linx almost every time on first boot after raising clocks regardless of vcore. But if I just leave it there anyway (as outside of linx, bc2 and other stuff is perfectly stable) and later reboot like normal pc usage rebooting the next day or something. Linx never fails.

Right now I'm linx 20 stable with much lower vcore/vtt than usual at 4.4ghz only thing different is that these bios settings have been set for a few reboots now. Seems good news as it means I may be able to get to my original goal of 4.5ghz 24/7 with all voltages under intel max.

[mike]

Interesting guide thanks ct, on a related note I fail linx almost every time on first boot after raising clocks regardless of vcore. But if I just leave it there anyway (as outside of linx, bc2 and other stuff is perfectly stable) and later reboot like normal pc usage rebooting the next day or something. Linx never fails.

Right now I'm linx 20 stable with much lower vcore/vtt than usual at 4.4ghz only thing different is that these bios settings have been set for a few reboots now. Seems good news as it means I may be able to get to my original goal of 4.5ghz 24/7 with all voltages under intel max.

It's not too uncommon that people overdo it with voltages...I always ty to keep every voltage as low as possible ( but I used to be the other way around....)

@ Mike - thanks, will do, question - where do you like to mount your probe - just beside the heat spreader?
Also, question - cpupll intel "absolute max" is 2v, why drop to 1.7 when 1.9 is < 2v - lower temps?

The lower your temps are, the lower the PLL voltage should be


Probe - yes - have it touch the IHS ( heatspreader ) on the outside