Hello, faithful readers! Don't worry, we're not dead, just busy half of the time and grabbing some summer sun during the other half :)
I just had an opportunity to test a Kingston triple channel DDR-2250 memory kit. Rated frequency is very high for a triple channel kit, the kit is clearly targeted at Gulftowns since 45nm Core i7 won't consistently reach that frequency, and certainly not on air.
Kit comes in a big box containing three modules, HyperX fan and some boring leaflets:

(click pics to enlarge)
I just had an opportunity to test a Kingston triple channel DDR-2250 memory kit. Rated frequency is very high for a triple channel kit, the kit is clearly targeted at Gulftowns since 45nm Core i7 won't consistently reach that frequency, and certainly not on air.
Kit comes in a big box containing three modules, HyperX fan and some boring leaflets:

(click pics to enlarge)
20/06: Some seven (GHz CPU)
Old-scool benching in last week (we tried our new dewar). We tried to push some of ours 775 CPU's to reach max validations. We tested about 15 CPUs. Main motherboard was ASUS P5B Premium (volt modded), ASUS P5B Vanilla (modded) and ASUS Stiker II Extreme (not modded). P5B Vanilla was not good in low subzero temps. It works in -180 degrees only a few minutes, then it was very unstable. So we change them to ASUS Stiker II Extreme, which is great mobo. Only cons is not working ClockGen, so we have to boot with benching frequency.
Anyway - we found a five 7 GHz CPUs.
Processors:

Our platforms:


And some scores:





We need to found solution for unstability with 1,95+ V on ASUS P5B Premium (maybe OVP voltmod?). When we do working 2,1 V we'll do another try (for higher CPU-z and some more bench). Almost 120l of LN2 and hours of hard work (and fun :) )
Thanks to GOODRAM (for great memory).
Anyway - we found a five 7 GHz CPUs.
Processors:

Our platforms:


And some scores:





We need to found solution for unstability with 1,95+ V on ASUS P5B Premium (maybe OVP voltmod?). When we do working 2,1 V we'll do another try (for higher CPU-z and some more bench). Almost 120l of LN2 and hours of hard work (and fun :) )
Thanks to GOODRAM (for great memory).
Category: General
Posted by: yotomeczek
Our second AMD Sempron 140 was the lucky one - unlocked second core worked fine. Because of it's pure OC, we concentrate on benchmark wPrime with Sempron 140 in 2 cores configuration. Not impressive OC - but not bad efficiency from MHz :). We used GOODRAM memory GP1200D (rated DDR2 1200 CL5, 2,1-2,4 V) and Gigabyte motherboard GA-MA790X-UD3P. This motherboard is great for high HTT (bus).
With only 6 liters of LN2 we did 2 first place on HWBot ranking with 2 cores Sempron 140.

Benching Sempron 140

wPrime 1024M (more detail HWBot link)

wPrime 32M (more detail HWBot link)

First GOLD

Second GOLD
Thanks to Gigabyte (for lending us this great motherboard) and GOODRAM (for lending us great DDR2 memory)
With only 6 liters of LN2 we did 2 first place on HWBot ranking with 2 cores Sempron 140.

Benching Sempron 140

wPrime 1024M (more detail HWBot link)

wPrime 32M (more detail HWBot link)

First GOLD

Second GOLD
Thanks to Gigabyte (for lending us this great motherboard) and GOODRAM (for lending us great DDR2 memory)
15/06: Our First 400k+ Aquamark 3
So new dewar wasn't wait too long for first bench session.
We freez Gulftown (crappy one) and try to push some numbers in Aquamark3.
GeForce GTX 480 was quick enough even without extreme cooling.
So ... after a few try we sow 400k+ score :)
Good enough for us and this not so good Gulf.

400k+ done

platform
We freez Gulftown (crappy one) and try to push some numbers in Aquamark3.
GeForce GTX 480 was quick enough even without extreme cooling.
So ... after a few try we sow 400k+ score :)
Good enough for us and this not so good Gulf.

400k+ done

platform
15/06: New stuff arrived :)
So we've got it :)
It takes long, but finally there it is:

Bruker Baltic SK-40
As the name says - it's 40 liters capacity
We tried it for a first time and it's good
It takes long, but finally there it is:

Bruker Baltic SK-40
As the name says - it's 40 liters capacity
We tried it for a first time and it's good
In our hands just dropped 4 GTX480 and a monster motherboard - Gigabyte X58 UD9. UD9 is a huge mobo. It's needed to have 11 slots case for them. Thanks to two NVIDIA NF200 it is possible to build 4 way SLI graphics system. Fermi needs a lot of power, but we've got excellence PSU - Silverstone ST1500. This PSU have everything we needed for 4-way SLI, including two CPU 12V connectors. System worked fine, and with 5 GHz Intel core i7-980X CPU system needed about 1500W peak power from the wall (PSU give about 1200W to hardware). As you see with ST1500 we have some room for harder OC.

Vantage score with 4 Fermis (more detail HERE)

Platform

Perfect PSU :)

4xSLI ready :)

A lot of amps here
It was only short time experiment, so we did only a few benchmarks (not all success with 4-way SLI).
We would like to thank:
Gigabyte (UD9 & GTX 480), Silverstone (ST1500), Zotac (GTX 480), MSI (GTX 480), Palit (GTX 480)
We are hoping that there will be more scores to come :)

Vantage score with 4 Fermis (more detail HERE)

Platform

Perfect PSU :)

4xSLI ready :)

A lot of amps here
It was only short time experiment, so we did only a few benchmarks (not all success with 4-way SLI).
We would like to thank:
Gigabyte (UD9 & GTX 480), Silverstone (ST1500), Zotac (GTX 480), MSI (GTX 480), Palit (GTX 480)
We are hoping that there will be more scores to come :)
Final scores for Polish qualifiers GOOC 2010:

It seems that performance per clock was quite good for us, except Pifast :)

It seems that performance per clock was quite good for us, except Pifast :)
22/05: GOOC 2010 Polish finals
Hello there,
Yesterday we had a pleasure to participate in Polish qualifiers to Gigabyte Open Overclocking Championship. Four teams were invited: Ryba and Chaos, Devilek and Kate, XtremeAddict and Mareczq, and us. All teams used a Core i7-980X CPU on Gigabyte X58A-UD3R motherboard, Kingston HyperX 2000 CL9 memory, Kingston 128 GB SSD and CoolerMaster RS600 PSU. We had a bad luck in CPU drawing - our chip had coldbug at -90°C, while other teams benched at around -120°. We did all we could with such bad CPU and reached the following scores:
(click pics for full size)

14.84 sec in Pifast - 4th place

3 min 00.828 sec in SuperPi 16M - 3rd place

3.281 sec in Wprime 32M - 2nd place
This allowed us to take 3rd place overall. Last round (Wprime 32M) was very interesting - with Ryba and Chaos in the lead and XtremeAddict a few points behind, our Wprime score was about to decide who will win the trip to European GOOC finals. We managed to take 2nd place in Wprime, putting XtremeAddict on 3rd place (2nd overall) and securing Ryba's win.
Gigabyte Poland did a very good job organizing the event, except that only selected teams were invited to take part - taking the "Open" out of "GOOC". It would also be nice if the chips were pre-selected by Gigabyte's in-house overclockers to be at least roughly equal in temperature tolerance. It was a nice opportunity to meet other clockers and flex our oc-ing muscles :P We hope next years GOOC will be even better (and that we finally find a good chip :D ).
Congrats to the winners and thanks to Gigabyte for inviting us. We'd also like to thank Terra.com for lending us an X58A-UD3R motherboard for practicing.
Yesterday we had a pleasure to participate in Polish qualifiers to Gigabyte Open Overclocking Championship. Four teams were invited: Ryba and Chaos, Devilek and Kate, XtremeAddict and Mareczq, and us. All teams used a Core i7-980X CPU on Gigabyte X58A-UD3R motherboard, Kingston HyperX 2000 CL9 memory, Kingston 128 GB SSD and CoolerMaster RS600 PSU. We had a bad luck in CPU drawing - our chip had coldbug at -90°C, while other teams benched at around -120°. We did all we could with such bad CPU and reached the following scores:
(click pics for full size)

14.84 sec in Pifast - 4th place

3 min 00.828 sec in SuperPi 16M - 3rd place

3.281 sec in Wprime 32M - 2nd place
This allowed us to take 3rd place overall. Last round (Wprime 32M) was very interesting - with Ryba and Chaos in the lead and XtremeAddict a few points behind, our Wprime score was about to decide who will win the trip to European GOOC finals. We managed to take 2nd place in Wprime, putting XtremeAddict on 3rd place (2nd overall) and securing Ryba's win.
Gigabyte Poland did a very good job organizing the event, except that only selected teams were invited to take part - taking the "Open" out of "GOOC". It would also be nice if the chips were pre-selected by Gigabyte's in-house overclockers to be at least roughly equal in temperature tolerance. It was a nice opportunity to meet other clockers and flex our oc-ing muscles :P We hope next years GOOC will be even better (and that we finally find a good chip :D ).
Congrats to the winners and thanks to Gigabyte for inviting us. We'd also like to thank Terra.com for lending us an X58A-UD3R motherboard for practicing.
I just tried our trusty old GoodRAMs PRO (Elpida Hyper, one of the first batches) on ASUS Crosshair IV Formula.
It was surprisingly easy to run four sticks (4x 2 GB) at DDR-1800 with pretty tight timings: 6-7-6-20 1T:


(click for full size)
With another CPU and new Crosshair bios another 50 MHz should be possible. More testing to follow :)
It was surprisingly easy to run four sticks (4x 2 GB) at DDR-1800 with pretty tight timings: 6-7-6-20 1T:


(click for full size)
With another CPU and new Crosshair bios another 50 MHz should be possible. More testing to follow :)
Nasi przyjaciele z Corsaira byli tak mili, że przysłali nam swój najlepszy zestaw dwukanałowych pamięci - Dominator GTX1.
Przy domyślnym taktowaniu 2400 MHz nie mogliśmy ich testować w zwyczajny, nudny sposób - sprawdziliśmy jakie najwyższe taktowanie uda się osiągnąć chłodząc pamięci suchym lodem.
Wykorzystana platforma:
Po zaizolowaniu okolic gniazd pamięci przeciw wilgoci wykonaliśmy z folii aluminiowej rynienkę między modułami, w do której wlaliśmy aceton i wsypaliśmy suchy lód:

Wyniki i więcej szczegółów znajdziecie na forum XtremeSystems lub, jeśli wolicie po polsku, PurePC.
Przy domyślnym taktowaniu 2400 MHz nie mogliśmy ich testować w zwyczajny, nudny sposób - sprawdziliśmy jakie najwyższe taktowanie uda się osiągnąć chłodząc pamięci suchym lodem.
Wykorzystana platforma:
- Procesor Core i7-860 - dziękujemy firmie Terra.Com za wypożyczenie
(chłodzony agregatem jednostopniowym do -45°C) - Płyta główna ASUS Maximus III Formula
- karta graficzna HIS Radeon HD5770
- Zasilacz Enermax Revolution85+ 850 W
Po zaizolowaniu okolic gniazd pamięci przeciw wilgoci wykonaliśmy z folii aluminiowej rynienkę między modułami, w do której wlaliśmy aceton i wsypaliśmy suchy lód:

Wyniki i więcej szczegółów znajdziecie na forum XtremeSystems lub, jeśli wolicie po polsku, PurePC.
