Oj0's Old Skool Benching, The Type That He Lives On, December 2007 -

[Oj0]

28 December 2007 - 09:48 AM

Athlon64 3700+ @ 3103MHz 1.525v AIR


DFI LP UT CFX3200-DR @ 411MHz HT


2x256MB Kingston HyperX BH5 @ 290MHz 2-2-2-5 3.94v Dual Channel


2x256MB G.Skill TCCD @ 350MHz 3-4-4-8 3.1v Dual Channel

30 December 2007 - 03:03 PM

Firstly, an UPDATE!!


3182MHz @ 1.55v VID Special @ 107.2 or 107.4% (cna't remember which it is) on STOCK AIR COOLING with ghetto mounting (the HSF is balancing on the chip) and Zalman grease

30 December 2007 - 10:58 PM

Some more TCCD results


2x256MB G.Skill TCCD @ 337MHz 2.5-4-4-8 3.0v Dual Channel

31 December 2007 - 05:31 PM

Another update:


TCCD @ 332MHz 2.5-3-3-7


Max CAS 2.5: TCCD @ 342MHz 2.5-4-4-8

To sum it up so far, here are the best results again:

Athlon64 3700+ @ 3182MHz on AIR
411MHz HT
2x256MB BH-5 @ 290MHz 2-2-2-5
TCCD @ DDR700
TCCD @ 332MHz 2.5-3-3-7
TCCD @ 342MHz 2.5-4-4-8

All results are dual channel

03 January 2008 - 04:12 PM

SuperPi @ 3.04GHz


Priming @ 2774MHz @ 1.328v


Priming @ 2852MHz @ 1.376v

04 February 2008 - 04:48 PM

DFI LP UT NF4 Ultra-D @ 420MHz

09 February 2008 - 10:01 AM

DFI LP UT NF4 Ultra-D @ 425MHz


Athlon64 FX-55 @ 3154MHz 1.55v

I have validations up to 3177MHz but they're all corrupt

I can post all the way up to the maximum 456MHz HT, but it won't load Windows. It sits at the screen where it checks the backup CMOS - it's still responsive in as much as you can switch Numlock, etc on and off and CTRL-ALT-DEL restart, but it doesn't do anything more.

[Oj0]

13 April 2008 - 11:09 PM

So I decided to hit the TCCD again this evening. I decided to use the Ultra-D for a change as it seems to be the board for TCCD. I started fiddling and within 5 minutes I was posting at DDR720, but that's as far as it went. Eventually I managed to get Memtest to open at a gargantuan 3.5VDD, so I decided tonight would be the night that I finally break my CPU-Z plateau. A Windows installation, BIOS flash and another hour later I still couldn't validate anything higher than 345MHz (DDR690), so I decided to call it a night.

I was still in the mood to fiddle around a bit, so I decided to try my faithful FX on the new BIOS (623-3 - Thanks B)). My previous best was 3154MHz at 1.612v. Drum roll please:

Athlon64 FX-55 @ 3189MHz 1.52v
Athlon64 FX-55 @ 3205MHz 1.616v

I FINALLY BROKE 3.2GHz ON STOCK AIR COOLING!!!!

Oh my word, what a difference, and all that from a BIOS change! 45MHz increase at 0.092v less and 51MHz to my maximum validation. I could save validations all the way up to 3230MHz, but they're corrupt. As you can see, it's not scaling to well with voltages anymore, a 0.096v increase only gained me 16MHz, so I decided not to increase the voltage any further.

I'm thinking it's about time to put my 3700+ on the Ultra-D... Oh yes, I've got a 4000+ on its way to me that's even better than the one I posted above - how does 3.1GHz STABLE at stock voltage sound? :P

That's it from me for tonight, next up will either be 3250MHz validated on stock air or I'll be on phase change

29 April 2008 - 12:18 PM

OMW, why can't every board be like this??? I set 1.456v (add or subtract a few millivolts) in the BIOS, and guess what I got? 1.455v idle, 1.455v load, 0.00v ripple, confirmed with a DMM! Not even my other DFIs could match this (LP UT NF4 SLI-DR Expert, LP UT NF4 Ultra-D, LP UT NF4 SLI-DR).

11 May 2008 - 05:20 PM

I'm aiming for sub 30 seconds at 2.7GHz, so far I've come up with this:

Not bad for 245MHz 2-2-2-5, eh??? This is just to show what tweaks can do. 270x10 2-2-2-5 should put me well below the 30s mark. My efficiency is 81650, I'm aiming for sub 80k next

17 May 2008 - 08:23 PM

270x10=2697MHz 2.0-2-2-5, efficiency 80,659 (previous best 81,650 at same speed)

Next goal: 29.663s for an efficiency of 80,000

EDIT: I also saw 25MB RAM used just before that run

18 May 2008 - 09:28 AM

DisablePagaingExecutive = "1"
SecondLevelDataCache = "1"
LargeSystemCache = "1"

Copy Waza > Kill Explorer > Pi > Open Explorer > Pi = best run

29.875s @ 2697MHz = efficiency of 80,573



RAM usage at 28,416KB - 3,036KB (Task Manager) = Windows idling @ 25,380KB, 24.79MB

06 July 2008 - 05:39 PM

So I used a bit of LN2 to keep my chip around @ -40c (thanks Marc and Seth for pouring), and 1.82v later I hit this:


Athlon64 FX-55 San Diego @ 3916.49MHz
Third on hwbot.org

Seth and I took it to about -45c core and JUST under 2v. The yesterday mourning I felt like really pushing for the limit so I started pouring it like water: -50c to -80c wouldn't pass 3.6GHz, -166c block started off with more than a 100c delta between the core and block and eventually got me -90c odd on the core with 2v got me 394xMHz but it froze while saving validation and wouldn't boot again. If you don't kill a chip you haven't found its true limit

Hardware used:
Athlon64 FX-55 San Diego CABCE 0516WPMW
DFI LanParty UT NF4 Ultra-D
1*256MB Mushkin Blackline PC3500 BH-5 @ 217MHz 2-3-2-5 2.9v
Radeon X1950XT
Random 500w PSU which I prefer to my AOpen 500w

EDIT: Thanks to Seth for this picture:


19 July 2008 - 08:58 PM


AthlonXP 2600+ @ 2542MHz AIR


AthlonXP 2600+ @ 265MHz FSB all AIR

Keep in mind that this is an AthlonXP non -M, i.e. locked multiplier.
Sixth on hwbot.org

08 August 2008 - 10:47 PM

Vapochill time. I'm using the Intel boot as it lines up perfectly with the core and I couldn't be bothered to change them :P I don't have any method of securing the head right now other than... A shot from the top:


And the head attached:


I didn't get any core contact whatsoever, within seconds the CPU hit 108'c and the PC shut down. Here's the second mount:


Finally, the whole setup ready to go:


Unfortunately, couldn't get a tight enough mount using string so tomorrow I'll bring the cable ties. While I didn't actually achieve anything, I had loads of fun

11 August 2008 - 09:40 PM

Three 3700+s later and here I am. Well this is 18th on hwbot and nowhere near good enough, but it's about the best I could get out of the cooling I had available and two evenings.


Athlon64 3700+ @ 3446MHz 1.82v Vapochill LS @ ~ 10'c

LN2 is going to help a lot this weekend.

EDIT: VERY nice cold bug, I could do 380MHz at -10'c

13 August 2008 - 09:22 PM

Mini update: Added 9MHz

Athlon64 3700+ @ 3455MHz 1.92v Vapochill LS @ ~ 45'c

I'm badly temperature limited now, pictures coming tomorrow.

The bloody Vapochill sounded like a running tap or boiling kettle, I gave her a nice workout :P

[Oj0]

14 August 2008 - 09:29 AM

BIOS shot of the voltage


Shot of the voltage and temperature. Ambient was 16'c, so I'm above ambient on phase change The Vapo dies at about 150w load

22 August 2008 - 09:17 AM

New World Record - Athlon64 X2 6000+

ripping.org
hwbot
CPU-z

M.I.H and S.I.H

12 November 2008 - 02:04 PM

600MHz 3-3-3-9

13 November 2008 - 08:49 AM

Much better PiFast - it didn't give errors this time

15 November 2008 - 12:28 PM

I used a motherboard with only one surviving memory slot by accident

Second place wPrime 32M


First place wPrime 1024M


Not bad considering the RAM I was using :P

I had a 32m Pi at 282MHz 2-2-2-5-1T but the time stank (it needs 2x512MB).

[Karmakazi]

WOW Jonathan this is an Epic Thread

I take it you found your flash drive??? (I hope)

[Oj0]

No, check the dates... This thread was four pages long with almost 3x the content but I removed the stuff I felt wasn't necessary. Core 2 and AM2 clocking, etc.

Oh, and THANKS

[knopflerbruce]

Looks nice K8 FTW!

[Oj0]

I only had one stick of RAM to use for these, but here's a 4002MHz 1M and 3937MHz 32M: