3000+ venice stuck @ 2.52

My venice is running at 2.52GHz 6:5 divider at 1.52v and Prime95 stable for almost 7 hours (choose to end test). I'm running my memory at 2.5-3-3-5 @ 2.85v. When I try to do 2.61 @ 1.52v it fails Prime95 fairly quick and it takes 1.62v to run for however long w/o any errors and memory at 2.5-4-4-7 2.85v(rated timings). 9x300 @ any voltage will not even post. If I run the 133mhz memory divider I can load windows @ 9x300 1.62v, but it fails Prime95 almost instant. I've tried relaxing my timings a little bit as that didn't help. I also have ran CPU Burn-in twice for 8hours and 9hours with the processor running 2.52GHz @ stock vcore. Should I run cpu burn-in for ~18 hours in hope to break this plateau or is my venice just not the badass OC'er I was hoping it would be. I'm also contemplating about dropping this stock cooler and getting an XP-90 and seeing if that won't improve my OC. I'm also thinking maybe my psu is the problem since I noticed my 6800 will barely OC above stock settings unlike before on my athlon xp setup. Any advice is appreciated.

System Specs:
A64 3000+ venice 2.52GHz 1.52v
1GB GeiL DDR500 @ 2.5-3-3-5 2.85v
MSI Neo2 Platinum (1.8 Bios)
eVGA GeForce6800 (12p/6v)
WD 120GB
Enlight 420watt (only 18a on 12v)

Please use paragraphs.

From your system specs, I would almost garuntee that your power supply is the bottleneck.

Still considering that it is a 3000+ decent OC and it may be that you have just reached the physical limit of the chip ...... You could invest in either an antec or a thermaltake purepower... I have a 550 and it is extremely stable at the rails...for half the price of antecs etc...

Yes, 18a on the 12V is pretty bad. You could run a burn in if you get a new PSU and it could help slightly but I think getting a new PSU with ~30 A on the 12V rail and at least the same wattage as your previous one should help a lot.

By the way, a 700mhz overclock on the 3000+ Venice which runs at 1800MHz is a pretty good one, especially on stock cooling!!

IMHO I would dare you might not get anything further from the 3000 without phasechange or some other kind of extreme cooling and maybe not even then ....I mean a 30% OC is not bad on any chip....

Golflad, remember this is a Venice and they do overclock well. However I have heard that the general overclock on good air is about 500MHz for the 3000/3200/3500.

Alright thanks for all the input guys and I apologize for my poor paragraph structure, I had a tee time to catch. I am more than pleased to have a 700MHz OC seeing how some people aren't as fortunate; but I want to get the maximum potential out of this processor. I'm gonna be looking for a new psu now and any suggestions would be helpful. Thanks

yeah, take a look at a new psu and an upgraded fan/heatsink.

If higher OC is an important goal and budget is not an issue, I would get a XP-90 (~ $30) especially if it is summer time for you.

For PSU, may not just for this setup, but also for future systems such as PCI-e, NF4, SLI, .... I would either get a Fortron AX500A (15A on 12V1 and 15A on 12V2, total 30A) or an Antec True Power II 550 (19A on 12V1, 19A on 12V2, total 38A). Both can be had for under $100. They both have native 24-pin power connector and can be used for motherboards with 20- and 24-pin (24-pin for most PCI-e and NF4 board) power connector. For the Antec, make sure it is the new version II which supports SLI, with two PCI-e graphic connectors, .... The Antec TP II 550 is plenty for growth.

Your current MSI Neo uses 20-pin connector, the Fortron AX500A has 24-20 adapter and the Antec has direct, deteachable 20-pin power connector to fit the MSI Neo board.


PSU rating estimate for some 939 CPU and system


For Venice, from the data I collected and analysed,

overclock_frequency_average = 2781 MHz
overclock_frequency_standard_deviation = 141 MHz
overclock_voltage_average = 1.548 V
overclock_voltage_standard_deviation = 0.102 V

- Venice/SanDiego can be clocked higher, about 200 MHz at least on average than Winchester, NewCastle, ClawHammer.
- The 90 nm SOI with DSL strained silicon (e.g. SanDiego/Venice) further improves the frequencies (by 200+ MHz on the average) compared to 90 nm SOI (non DSL) (e.g. Winchester).

Overclocking frequecy and voltage of various A64 (post 114)

Do "best" CPU voltage and frequency exist for overclocking (post 115)



Cool and Quiet

Also for 24/7 running of the system, you may consider setting up Cool and Quiet. It requires some works to get it right, but the system would be able to run much quieter and cooler. During benchmarking, you may disable the CnQ feature in the bios.

Cool and Quiet (CnQ) Testing
Running the computer Cool and Quiet
Cool and Quiet with new higher overclocking (continued)

hey i was planning to overclock my 3000+ too, one question did you get to 225 fps without any problems? nice overclock by the way, i will be doing mine when i get a new power supply

oh and wats a burn in?

I did 225fsb easily with no problems at stock vcore. I had 225fsb runnin stable back when I had GeiL value series ddr400 memory. Also what's the difference between dual rail and single rail 12v psu's. I still haven't figured that out.

I plan on getting the Fortron AX500A for ~$78 and the Freezer64. With the Fortron how do I combine the 12v rails to make it one 30a 12v rail.

i think one for motherboard and other for something else

Yoshpop said:

I plan on getting the Fortron AX500A for ~$78 and the Freezer64. With the Fortron how do I combine the 12v rails to make it one 30a 12v rail.

Click to expand...

For those PSU's with dual 12V output such as the Fortron Blue Storm, the Antec True Power II, ..., you don't combine the two 12V outputs. They are designed so that one 12V output is for powering the CPU via the 4-pin connector, and the other 12V output is for the rest of the system (video card(s), hard and optical drives, fans, ...). This way, hopefully, the voltage fluctuations and transients caused by the load change of the CPU and that caused by the load change of the video card(s) and hard drives (during spin up) would affect one another to a much lesser extent than if all the compoents are on the same 12V output.

Your cpu is maxed out and thats an above average overclock. I have the exact same one and it maxes out at 2.45GHz on 1.52v tried 20MHz higher and it crashed but is rock stable at 2.45. I wouldnt complain.........

so you havent run the 18hour burn in? chances are thats what is limiting your OC.. i did an overnight (so about 12hours) of burn in using SETI@home, and I am happily running rock stable @ 2.7ghz on stock cooling.. but my PSU is only a 400W with 18a on the 12V, and still no problems.. so im going to guess that its the burn in.. if not, then Im not sure sorry
Careface*

Yes, that's a 40% overclock and nothing to sneeze at. (My Winchester will do that on stock voltage but that's just dumb luck on my part) By all means get that new cooler on there as that has to be a strain on stock cooling. A64's don't heat up much with an overclock till you start hitting the voltage and then they go up pretty quickly.
Adriana, a burn in is basically a means of breaking in hardware so that you get maximum performance from it. There is a freeware program, CPU burn-in, which many people around here use. There is other software that can be used for burn in as well.

yoshpop how did u do a 6:5 divider on the board u have?

icemanjs said:

yoshpop how did u do a 6:5 divider on the board u have?

Click to expand...

I can go to 9x282 at 2-3-3-6, vcore at 1.42v. I assume it would go higher with more juice but I only recently upgraded to the XP90 from a SwiftTech so I haven't really tried it yet. (The AS5 is still breaking in)