• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

AC's Phenom Review and Techinical info

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Ac sounds like you need to spend the money to get a bad *** mobo that will let you change stuff more. it looks like with your low Vcore your doing ok.
how fast do you have the memory running and hows the temps under load?
 
What kinda bandwidth are you looking for on the phenoms?

just wondering as I was very pleased to see 11k mbs in sandra.
I haven't tested my rigs for bandwidth as this point so I'll do that this evening to see what's really going on with the AM2 rig and the 165's. Sandra shows my Opty 180 (246x11 @ 1:1 CAS 2.5) @ 6691/6662 and 69ns latency and that's one of my "slow" RAM rigs. To match that I'd need a little over 13k on a quad.

The Opty 165's are running 263 & 256 MHz RAM (CAS 3 & CAS 2.5, respectively). I'll test those tonight as well as the AM2 rig but I'll have to use Memtest on the AM2 rig, which is now Linux ...
 
Last edited:
@QuietIce- the difference in SETI times may be due to the lack of L2 on your AM2 vs. your 939 Opterons. We see this on the Folding side of the fence as well. The larger L2 really makes a big difference.

I have a 170 @ 2.7GHz & a 5000 BE @ 3.1GHz both Folding in Linux and the 5000 BE just barely out paces the 170.
 
Ok CPU multi not available and I'm stuck at 11 but I'm playing with more OCs.
Clock is at 221, CPU 2431 VCore still at 1.265 and NB speed is still 1990 or just under 2000.
I might settle for 2450 but wish I could drop the NB speed to be sure.
The testing goes on.



I noticed that there is 2 ways you really want to go about clocking the setup

which is either NB and HT = same speed or speed difference between cpu and nb speed = difference between nb and ht link speed.



I managed 12 mins per frame smp last night at 10x258 with ram over 1k and nb over 2k ht around 1500. This is on a stock cooler BTW





To match that I'd need a little over 13k on a quad.


I think I can manage 12-13K if I where to up my ram voltage and try for it but atm I get around 11k with ram at 1k ddr and all timings set to auto....
 
My RMAed GA790FX comes in today so I may swap back if I can find people with 9500s running more than 2432. I tried 225x11 and it crashed about 10 mins later. 220-222X11 it's stable for at least 30 mins.
Looks over at an old school 4600X2 doing 2750 in a $50 board :) Looks at Phenom doing 2432 :( but the Phenom rips that 4600 a bad new arshole so the Phenom still has power.
 
I managed 12 mins per frame smp last night at 10x258 with ram over 1k and nb over 2k ht around 1500.

I just ran an SMP unit for folding. Both to confirm stability and also to give you guys more data points.

At 20% it was running 12:30 frames, at 50% and above it started running 12:45 frames. Memory slowing me down I guess.

So far I'm 12 hours prime stable, over 24 hours of SETI with no validation errors and 1 SMP unit returned ok. Not bad as a stability test I guess.
 
I just ran an SMP unit for folding. Both to confirm stability and also to give you guys more data points.

At 20% it was running 12:30 frames, at 50% and above it started running 12:45 frames. Memory slowing me down I guess.

So far I'm 12 hours prime stable, over 24 hours of SETI with no validation errors and 1 SMP unit returned ok. Not bad as a stability test I guess.
SETI is a GREAT stability test! Nowadays I run Orthos for about an hour then put the rig on SETI. After you've crunched 2-3 days than check for compute errors - you'll have enough results by then (IMO) to confirm stability. :)

SETI CPU times as follows:
CPU / speed @ RAMspeed ................................ 50-52 Cr; ~54 Cr; 73-75 Cr (credits)
FX-55 (90nm) 2.9 GHz @ 223 MHz CAS 2.5: 76-7800; 82-8400; 88-9000
Opty 165 2.9 GHz @ 263 MHz CAS 3: ................. 78-7900; 85-8600; 90-9200
5000 BE 3.2 GHz @ 400 MHz CAS 4: ................... 80-8300; 85-8800; 93-9500
4400 X2 2.6 GHz @ 238 MHz CAS 3: ................... 85-8600; 90-9300; 9700 (only found one)
Your Phenom (for easy comparison):
Phenom 2.6 GHz @ 315 MHz CAS 5: ................. --------; 82-8300; 8800 (72 credit)

You're getting about the same times as my best cruncher but you're running 10% slower on the CPU!
I'd say that's a MAJOR step up!!! :):thup:

Now if AMD can get those Phenoms running just a little faster I'd say we have a winner! ;)
But even now it might be cheaper to build a Phenom system than a pair of Opty 165s - I'll have to do some cypherin'.

@QuietIce- the difference in SETI times may be due to the lack of L2 on your AM2 vs. your 939 Opterons. We see this on the Folding side of the fence as well. The larger L2 really makes a big difference.

I have a 170 @ 2.7GHz & a 5000 BE @ 3.1GHz both Folding in Linux and the 5000 BE just barely out paces the 170.
I hadn't thought of that - DUH! Now that I think of it ALL my s939 rigs are 2x1Mb L2, even the X2's I have.
If anything the L3 on the Phenom should HELP, if true (and from the above results it might be!).

Guess I'll have to get a rig built with this (s939) 4200 I have sitting around so I can compare - it's the only 512K chip I've bought ...
 
Last edited:
Hey guys question? Im looking at getting the 9600 BE w/GIGABYTE GA-MA790X-DS4 AM2+/AM2 AMD 790X ATX AMD Motherboard from the Egg ofc and was wondering would it be better to get a single stick of 2Gb DDR2 800 (PC2 6400) or 2x1Gb sticks? I'd more than likely be topping out at 4Gb down the road so it wouldn't matter since the board has 4 dimm slots. But from a proformance point of view would i see any gain on it one or the other way. Also it runs stock at DDR2 800 right?
 
I just ran an SMP unit for folding. Both to confirm stability and also to give you guys more data points.

At 20% it was running 12:30 frames, at 50% and above it started running 12:45 frames. Memory slowing me down I guess.

So far I'm 12 hours prime stable, over 24 hours of SETI with no validation errors and 1 SMP unit returned ok. Not bad as a stability test I guess.
Good progress. That means better ram will drop you somewhere in the 11s. This may be enough proof for me to jump back over to the GA790FX.

Hey guys question? Im looking at getting the 9600 BE w/GIGABYTE GA-MA790X-DS4 AM2+/AM2 AMD 790X ATX AMD Motherboard from the Egg ofc and was wondering would it be better to get a single stick of 2Gb DDR2 800 (PC2 6400) or 2x1Gb sticks? I'd more than likely be topping out at 4Gb down the road so it wouldn't matter since the board has 4 dimm slots. But from a proformance point of view would i see any gain on it one or the other way. Also it runs stock at DDR2 800 right?

You will want 2 sticks, ganded mode has proven to boost speed by up to 25%.
If you have to pick up some cheap 667 until you can swing for the 2x2G 8000-8500s. They are well worth it.
 
I posted a new screen shot and it's pretty nice really! LMK what you think god I can't wait to get my xp-90c back on this thing.
 
so iyo do you all think the phenom is better than the x2 6400 as far as a gamming build? i know it clocks slower but quad over dual....?:confused: Its the BE9600 with the Gigabyte board or Athlon 64 X2 6400+ Windsor with ASUS M2N-SLI Deluxe AM2 NVIDIA nForce 570. i thought i was planning the 6400 but you guys got me looking again. With the BE id be looking at about $80 more for the complete build.
 
Last edited:
For current games I'd say there is no benefit for having a quad over a dual. Even benchmarks of Crysis show no tangible speed benefits from the quad. You will almost certainly be GPU limited regardless of which CPU you choose.
 
Back