Benchies from my fastest FX

[Daveburt]

This was my first FX chip, week 35/2011. It's an 8120 (1135 BPA).
It's a low VID chip (1.2375v) it runs pretty hot on ambient temps, but it's actually my best clocker as long as I keep temps under control...
Some of these benchies have been posted on other forums, but since I'm pretty new here and alot of you folks seem to be interested in high clocks, I thought I'd share...

First off... I have above average water cooling and a couple of these took advantage of the weather, but there's no extreme (DICE/LN2) cooling used!

Here's a pic of my Rig:


My best CPU-Z valid was ~5.4Ghz (link in my sig). It was done by sitting my rig on the patio when it was 32f (0c) outside.
[http://valid.canardpc.com/show_oc.php?id=2091880

Best Cinebench 11.5 (cold air):


Here's a serious WinRAR (beta2) push using indoor ambient air (Xigmatek 1283 ~70f):


Finally... As worthless as Superpi is when judging modern processors here's some results using indoor temps (73f/Xigmatek air).
The Wprime score in this SS is pretty impressive IMHO, so I thought I'd include this shot (the WinRAR score is from the old version)...



I've seen a number of these same benchmarks on review sites that are WAY lower than what I've been able to achieve...

Granted, some of my ambient cooling solutions are mildly extreme, but it makes me wonder if some of these folks actually know how to tune FX chips...

 

[RGone]

Thanks there "Daveburt", some good numbers. Nice looks inside the case too. Have you tried to run P95 in blend mode for a couple of hours to see if those high speeds are stable? The reason I ask is there is a small rumor that the later FX-processor might run a little faster with less vovltage.

 

[Daveburt]

Thanks there "Daveburt", some good numbers. Nice looks inside the case too. Have you tried to run P95 in blend mode for a couple of hours to see if those high speeds are stable? The reason I ask is there is a small rumor that the later FX-processor might run a little faster with less vovltage.

Sup RGone...
I'll be the first to admit that P95 Blend (Custom/Max Mem) is the best way to test stability on FX chips!
I never run it for hours at a time though...

Here's a 45 minute Prime stability run on the 8120 with resonable v's:
This was stable for all my uses!



Nothing heats things up on FX like Prime does!
It's the best way to rough in stability, but I just don't like to push my gear that hard, and from my experience if you can get it Prime Custom stable >30 minutes it'll be good for daily use...

That 8120 is my fastest chip, but my 8150 is good for ~40Mhz more on ambient temps and generally runs cooler...
Yeah, those voltages where a little high, but I was pushing for max clocks!

Here's a Prime run on my 8150 (my 24/7 chip):


In all reality, they're pretty close in performance, but the 8150 has been more relialble...

And the geek in me like the 40Ghz boost (with stability)...

 

[keny]

Nice low vcore for a 5ghz+ chip

 

[(G{in}[AK)TION]]

if i had that setup, but with air cooling, i would use the FX outside during the winter. ^^

the case i would use would be the Antec Lanboy Air.

nice setup btw!

 

[RGone]

I agree with "keny", that appears a super chip and nice setting up that you did. Thanks man. RGone...ster.

Sup RGone...
I'll be the first to admit that P95 Blend (Custom/Max Mem) is the best way to test stability on FX chips!
I never run it for hours at a time though...

Here's a 45 minute Prime stability run on the 8120 with resonable v's:
This was stable for all my uses!



Nothing heats things up on FX like Prime does!
It's the best way to rough in stability, but I just don't like to push my gear that hard, and from my experience if you can get it Prime Custom stable >30 minutes it'll be good for daily use...

That 8120 is my fastest chip, but my 8150 is good for ~40Mhz more on ambient temps and generally runs cooler...
Yeah, those voltages where a little high, but I was pushing for max clocks!

Here's a Prime run on my 8150 (my 24/7 chip):


In all reality, they're pretty close in performance, but the 8150 has been more relialble...

And the geek in me like the 40Ghz boost (with stability)...

 

[Woomack]

@Daveburt , nice results for water cooling
My FX8120 has about the same results ( +/- voltage etc ) but I have never checked it on water.
Weird is that on cold water these cpus can make ~5.2-5.3GHz and on single stage only about 200-250MHz more. Simply they need LN2
On 1.55V this cpu can make some more than 5.38GHz just needs better cooling. Temps are killing results above 5.2GHz. 0.05V more and it will overheat, too low and it will crash.

And now funny thing ... my FX8120 is making ~5.5GHz on 1.55V and my i7 3770K is also making ~5.5GHz on 1.55V and the same cooling

 

[Daveburt]

Weird is that on cold water these cpus can make ~5.2-5.3GHz and on single stage only about 200-250MHz more. Simply they need LN2

Thanks for the info man...
I've actually been considering SS Phase or an A/C waterchiller to try and push this chip farther...
I know it has alot more in it (if I can keep it cool), but going through all that work for another 200 Mhz seems pretty useless. Return on investment and all...
I'm assuming you've tested yours on phase (seems similar to mine )?, if it's only good for another 1-200 Mhz it probably isn't worth the money...

I'm pretty sure this chip would shine on LN2 (7Ghz+), but I don't have a pot, and as much fun as it would be to play with extreme cold it's not really practical.

I'm buying a house next year and I'm going to try and find a place with a basement/garage. If I'm successful with that project it should give me more room to play with those options!

Living in an apartment kinda limits what I can do...

 

[Woomack]

I made some fast tests on air ( TT 120Ultra ) and later I moved to SS. Unit can make about -50*C idle but for this cpu it's too weak and under full load temp is nearly 0*C ( +/- 5*C ) so have to watch out for moisture etc.
Max was something near 6080MHz ( 2 cores and voltage ~1.63V ).

To be honest I was testing much more memory overclocking on this platform as cpu speed isn't anything special and without LN2 there is no fun.
Max memory clock was something near 1523MHz ( 3046 effective ) but anything above ~2200 is pointless for daily usage/benchmarks unless you set high CPU-NB clock.
Still nice platform to make max clock results
I haven't been testing it for over half year and my brother has now AMD rig. In last days I'm testing mainly P4

 

[marjamar]

I've played with A/C with my CHIVE and FX-8150 a month or so ago. I built a custom cardboard case and tried a few recirculating ideas. I was able to maintain a near 0 C. environment for the entire computer components without any moisture problems, but I didn't have anything but my Corsair H80 to cool the CPU which wasn't enough for more then a few minutes of overclocking at a time. I did get to 5.4 as I recall, but with no time to adjust or test much of anything, never got stable. I was stable at just over 5, but still no real time to test prime95 more then 5~10 minutes at a time.

The idea of air conditioning an entire computer still entices me. I have this idea (dream maybe) of building a fully air conditioned custom case where the air conditioner is within a pretty well soundproofed section below the computer hardware. The computer hardware would be more or less an open layout sort of fancy test bench style and have a full double-pane glass top. Keep the whole box air tight and filled with nitrogen gas to displace all air/moisture.

I wouldn't be surprised if I couldn't get ambient temps to about -10 C. and maybe reach a stable 5.7 or better, all cores running.

Anyways, it would be a fine piece of work to do, and very neat if it turned out.

Right now I'm waiting on a bunch of new stuff to make a new water cooled setup. More or less a tame rig, but nicer then I have now. I'll start my own thread when I start the new build. Most likely do a video for my youtube account as well.

Nice overclocking BTW!

-Rodger

 

[keny]

The only problem I can see from running a cooled air pc is that as soon as switch the ac off you are going to get moisture/condensation build up as the components will still be colder the air around them, you would need to run the ac 24/7

 

[marjamar]

The only problem I can see from running a cooled air pc is that as soon as switch the ac off you are going to get moisture/condensation build up as the components will still be colder the air around them, you would need to run the ac 24/7

I have alot of questions, but a few answers seem to filter through with my tests so far. I'm pretty sure the answer to running 24/7 is thermostat control over the compressor at a minimum, maybe the whole air conditioner. Also, dew can only occur when moisture is present in the atmosphere. Since I would have a totally closed environment for all of the computer and air conditioner evaporator, it will be filled with nitrogen gas, displacing all the normal air and moisture. It might be necessary to pull some vacuum to fully remove all moisture, but it would be essentially moisture free inside the top portion of the case.

We've done some of this, sort of, on some of our jobs. I own a sandblasting and coatings company and on occasion we do sandblasting on steel tank interiors. When the job length is more then a day or if we need to wait for some reason before applying coatings to tank interiors, we have flooded the insides of tanks with nitrogen gas, which will displace the air and moisture so no rusting will occur to the sandblasted surfaces.

I know, I know, not the same thing. But, I also know without trying something first hand, I seem to be left guessing what would have happened -- Which bugs me to death most of the time.

-Rodger