[xsuperbgx]

I have a complete AMD system with a slow dual core that has a hard time meeting SMP deadlines. The mootherboard is compatible with AM3 processors but only up to 95 watt (mATX). Can anyone give me ppd numbers on the phenom2 and on the athlon2 quad cores and maybe some opinions on which would be best.

Thanks....

[Bobnova]

I've had a PhII 810 cruising along at 3.3ghz making 5600ppd. They run nicely

[sneakysnowman]

My phenom II 940 @ 3.6GHz is producing 5500 ppd right now on a standard A2 linux 1920 point WU (project # 2671). It is bogged down slightly with 4 GPU clients

[chuckerants]

What CPU do you have?

Originally Posted by xsuperbgx

I have a complete AMD system with a slow dual core that has a hard time meeting SMP deadlines. The mootherboard is compatible with AM3 processors but only up to 95 watt (mATX). Can anyone give me ppd numbers on the phenom2 and on the athlon2 quad cores and maybe some opinions on which would be best.

Thanks....

[glussier]

Originally Posted by Bobnova

I've had a PhII 810 cruising along at 3.3ghz making 5600ppd. They run nicely

Can you post a FahMon screenshot showing that, Please? I'm sorry, but I don't believe 5600ppd at 3.3ghz for a Ph ii 810.

[Bobnova]

Can't, it's a part time thing. That's what it calculated to doing the math manually (compiling fahmon gives me headaches).
Consider sneakysnowman's 5500 at 3.6 while running multiple GPU clients.

[xsuperbgx]

I currently have the opteron 1210(1.8GHZ dual).

Does anyone have the athlon2 X4? I am mainly wondering how much difference the L3 cache makes. I was thinking the 620, or the 810 or 945. Cheaper would be better, but if the cache makes a big difference I may increase my budget.

[Bobnova]

mmm Dolk might know, seems like someone is bound to have fah benchmarked a 630 by now.

[Porvalsh]

I must be doing something wrong. I only get around 3000ppd with my Phenom II X4 955. I have it folding on a VM in Ubuntu 9.04.

I asked ICEBOB to run some benchmarks in folding with the Propus (Athlon II X4) but he hasn't done any. Maybe more people should ask him.

[glussier]

Originally Posted by Porvalsh

I must be doing something wrong. I only get around 3000ppd with my Phenom II X4 955. I have it folding on a VM in Ubuntu 9.04.

I asked ICEBOB to run some benchmarks in folding with the Propus (Athlon II X4) but he hasn't done any. Maybe more people should ask him.

If this is running at 3.9 on only 2 cores, this is around what you should get.

[dark bishop]

my phenom2 940 @ 3.5Ghz got 5200 ppd running a notfred vm and having the gpu fold.

currently on pause to go for 4.0Ghz

[Porvalsh]

Originally Posted by glussier

If this is running at 3.9 on only 2 cores, this is around what you should get.

Its VMWare Player 3.0, running all 4 cores though.

[glussier]

Originally Posted by Porvalsh

Its VMWare Player 3.0, running all 4 cores though.

Then, you are only getting around half of what you should get. Are all the cores busy @ 100%?

[Porvalsh]

I thought so but this weekend I can take a look at the setup again. Without PPD estimates under certain installations and clients its hard to tell if your numbers are good or not, lol. Being a new folder this has been the main issue. I used to only get like 1500 haha.

The CPU takes 9 minutes to complete 1%.

[ihrsetrdr]

Originally Posted by Porvalsh

I thought so but this weekend I can take a look at the setup again. Without PPD estimates under certain installations and clients its hard to tell if your numbers are good or not, lol. Being a new folder this has been the main issue. I used to only get like 1500 haha.

The CPU takes 9 minutes to complete 1%.

Post the Work Unit you are referring to[above] so we can do the math for the PPD.

I've strictly been on Intel for folding these past several years, but I'm interested in jumping on one of the AMD quads for a Linux SMP experience.

[Outback_Jon]

Originally Posted by xsuperbgx

I currently have the opteron 1210(1.8GHZ dual).

Does anyone have the athlon2 X4? I am mainly wondering how much difference the L3 cache makes. I was thinking the 620, or the 810 or 945. Cheaper would be better, but if the cache makes a big difference I may increase my budget.

Can't help exactly, but I just built two Phenom (not II) X3 8750 Black Edition systems. (Haven't started to overclock them yet, though) They run at 2.4GHz, while the socket 939 Opteron 165 (X2) is at 2.5GHz. (Yeah, the kids have better hardware at the moment) The PPD on similar projects is almost EXACTLY 1.5x what my dual-core Opty gets. Exactly as if I had a third core.

Now, the Phenoms only have 2MB of level three, but it really doesn't seem to have much effect. Maybe the bigger cache on the Phenom IIs would help more.

[Porvalsh]

Originally Posted by ihrsetrdr

Post the Work Unit you are referring to[above] so we can do the math for the PPD.

I've strictly been on Intel for folding these past several years, but I'm interested in jumping on one of the AMD quads for a Linux SMP experience.

The 8800GT is on another box.

[Porvalsh]

This weekend I moved my Radeon 4890 to another box. Now I'm running 3 GPU clients on my box and have seen my PPD on the proc climb to 4600, running at 3.5ghz. It seems that the ATI card was gobbling up clock cycles from the CPU? I didn't change anything with the VM client to see a bonus in PPD. All I did was uninstall the ATI drivers, and install the three Nvidia cards (8800GTS, 88800GTS, 8800GT).

Worth noting, now the ATI card is getting even worse PPD. Its running on a 939 Venice build and pulling in 2k ppd. Was getting a bit over 3k in the X4 rig. The 8800GTS running in the Venice rig was pulling 5600PPD. Just another reason why ATI sucks at folding I guess.

[Computekinc.us]

I just upgraded 4 dual core machines here with the 95w Phenom II 630 CPU's at stock clock (2.8Ghz) they pull 4200+ PPD. I am using notfreds under VMware workstation.

[Porvalsh]

Originally Posted by Computekinc.us

I just upgraded 4 dual core machines here with the 95w Phenom II 630 CPU's at stock clock (2.8Ghz) they pull 4200+ PPD. I am using notfreds under VMware workstation.

Do you mean Athlon II X4? I was not aware the Pii's came in a 630 marking. This sounds like the A2 X4 630 Propus chip.

If this is the case, then this is great news. 4200PPD for a $90 chip!

[Computekinc.us]

Opps, Athlon II is correct. I can't keep up with the name changes I guess.

[xsuperbgx]

So roughly, it seems that on the AMD quads, that someone should get 1500 or so points per GHZ. Also it seems that the L3 cache has a minimal effect on the ppd numbers.


Does that seem right?




Thanks for your replies!
It looks like an athlon2 620 may be in my future!

[onefstsnake]

My Phenom II x2 545 is getting ~2200ppd at the moment.
Stock clocked at 3.0ghz

Thats using VMplayer 3.0 running Ubuntu 9.10

[xsuperbgx]

My Phenom II x2 545 is getting ~2200ppd at the moment.
Stock clocked at 3.0ghz

Thats using VMplayer 3.0 running Ubuntu 9.10

That seems to fall in pretty close to 50% of what the quad does so that seems like fairly good numbers to me. But I may be oversimplifing......

[Computekinc.us]

Figure the Athlon II X4 630 to put out 3800 - 4200 depending on the configuration. I upgraded all of the dual core 90nm 125w cpus I had to these and doubled production and use less energy! The old Athlon 64 X2 6400+ CPU's are going for $100 on ebay currently - X4 630 @ newegg $122 = 1 happy camper right here!

[Bobnova]

2200 is about right in a VM.
Native linux with a 810@3.3-3.5ghz is around 5500, after you figure the loss for 2.10 on a dual core and the loss for a VM, 3ghz and 2200 sounds decent.

[Porvalsh]

Originally Posted by xsuperbgx

So roughly, it seems that on the AMD quads, that someone should get 1500 or so points per GHZ. Also it seems that the L3 cache has a minimal effect on the ppd numbers.


Does that seem right?




Thanks for your replies!
It looks like an athlon2 620 may be in my future!

Welll...seems like it has a decent effect. I mean, we're talking about a 1k PPD difference between Pii quad and the A II quad right?

Thats like 20%.

Overall though, not that big of a deal considering a 50 dollar graphics card can net you the same PPD as these quads.

EDIT:
Hmm, I'm getting 4600ppd on my quad with 3 GPU clients. Its running at 3.5ghz right now. If you're pulling close to 4000 @ 2.83ghz then thats pretty good. Not much difference...