i'm just curious about what average PPD people are getting on multiple platforms WITHOUT overclocking to see what platform might be worth putting money into...

really interested in:

Athlon IIs
Phenom IIs
Istanbul
Nehalem-EX Xeons
Core 2 E7000s, E8000s, Q8000s, Q9000s
i5 750, 800s/900s (with and without HT)
Clarksfield (if you're lucky enough to have one ;)

moderately curious about
Phenom Is (9000s as well as 9x50s)
Core 2 E4000s, E6000s

mention in the post if you crunch on a machine that's used for other things, too, at the same time, like browsing the internet, or if it runs uninterrupted.

---

i guess i'll offer mine:

http://boincstats.com/stats/host_graph.php?pr=rosetta&id=1147868

i7 920 (133 x 21 = 2.8ghz).. didn't really run long enough to establish a stable number of points per day, but ~3000 points +/- 20%, win7, hyperthreading on.

---

Here's the thing.

The latest AMDs are way cheaper than the latest Intels, BUT, AMDs Overclock like a charm and are rock solid.

Now, if you are looking for solid performance period. Then get Intel with Nehalem architecture. (Which also overclock like a charm, but go hard on the wallet).

And as a tip. Get an ASUS mobo.

I run both non-OC'ed and OC'ed machines.
All running Rosetta seamlessly.

---

well, i broke a $180 mobo when i broke my personal vow to never overclock... so yeah, not going to overclock ever again. plus the power it uses... i'm not sure if power use is linear to performance increase.

i'm actually quite happy undervolting my i7-860 to just a smidgen over 1.000v (and still prime95 stable)...

---

I'm only doing Einstein at his time so i cant contribute much here but still:

Core i7 920, stepping D0, great at stock and can go really high if you one day decide that overclocking is wise (coz it is), even without voltage increase D0 is great if you dont want to overvolt.... Overclockable i7 mainboards and DDR3 isnt expensive anymore and even cheaper if you are sure you will never OC. And dont listen to that crap lots of ppl like to repeat like parrots about "max 1.65v vDIMM", i never had my stepping C0 920 that low on the vDIMM and now its just a few weeks left till it been crunching various projects 24/7 for a year.

I guess i5 is nice too, but 2x2GB RAM isnt much for Rosetta, so for the memory its cheaper to go i7 with 3x2GB than 4x2GB for i5.

---

well, i broke a $180 mobo when i broke my personal vow to never overclock... so yeah, not going to overclock ever again. plus the power it uses... i'm not sure if power use is linear to performance increase.

i'm actually quite happy undervolting my i7-860 to just a smidgen over 1.000v (and still prime95 stable)...

"Over" volting is to make the CPU stable at higher clock rates. Increasing the volt alone won't make it any faster.

---

"Over" volting is to make the CPU stable at higher clock rates. Increasing the volt alone won't make it any faster.

my push from 3.6 to 3.7 might have been what killed it (plus lack of experience). i had to push it slightly above intel's recommended voltages. it's weird as it died a week or two after i'd pulled back to 3.1 and was at 1.25-1.29v.

anyways, what points per day are you getting, and at what speeds? (if any of you OCing readers are on R@H)

---

"Over" volting is to make the CPU stable at higher clock rates. Increasing the volt alone won't make it any faster.

my push from 3.6 to 3.7 might have been what killed it (plus lack of experience). i had to push it slightly above intel's recommended voltages. it's weird as it died a week or two after i'd pulled back to 3.1 and was at 1.25-1.29v.

anyways, what points per day are you getting, and at what speeds? (if any of you OCing readers are on R@H)

This PC: https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=1172219

Is overclocked from 2.7 to 3.1 GHz.

hasn't created any errors so far (except for the ones that fail right away)

---

New i5/750 build. Not overclocked (yet). Little over 1000ppd, 50/50 split between here and Einstein. 2 cores each. So it would be around 2000ppd if all 4 cores were here. Einstein is around 1500ppd (3000ppd if solo project). Both are close to leveling off, but not quite yet.

https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=1168293

---

Based on this chart, go with the AMD Phenom X4 905e, best bang for your buck compared to the AMD's that are that high on the chart.
http://boincstats.com/stats/host_cpu_stats.php?pr=rosetta&teamid=&st=0&or=10

It's an undervolted processor already, and has been tested to be stable at the lower voltages, and saves you some on electricity as well.
It goes for $209 on Compusa...
http://www.compusa.com/applications/SearchTools/item-details.asp?EdpNo=4695459

And $169 on Newegg...
http://www.newegg.com/Product/Product.aspx?Item=N82E16819103682

Looks like Newegg is the way to go for this one,and they have reviews on it too.

---

that chart is good, useful IN THOERY, but i question it's accuracy. i'm pretty sure i've gotten ~800-1200 points (at least) in a day on my X3 710 or 705e (and the chart even has the 720 below the 710)

---

Most any chart from boincstats will fail to account for any adjustment for resource share. In other words, it is going to show machines with 10% resource share to a given project and other machines with a 100% share all on the same chart. You have no way to know if the machine is overclocked. No way to know it is crunching other projects. No way to know if the amount of memory presently shown is what was in use at the time the stats were gathered. Was the machine busy with other work? Is the machine only powered on 8 hours a day? Does the machine crunch while in use by a user? Is BOINC allowed to use all of the CPUs at 100%? Is BOINC allowed to use enough memory that it can keep the CPUs busy the entire time BOINC is allowed to run? ...so some oddities are to be expected.

No fault of boincstats, it's just simply not easily extrapolated from the stats data provided by the standard BOINC server code.

---

Rosetta Moderator: Mod.Sense

Most any chart from boincstats will fail to account for any adjustment for resource share. In other words, it is going to show machines with 10% resource share to a given project and other machines with a 100% share all on the same chart. You have no way to know if the machine is overclocked. No way to know it is crunching other projects. No way to know if the amount of memory presently shown is what was in use at the time the stats were gathered. Was the machine busy with other work? Is the machine only powered on 8 hours a day? Does the machine crunch while in use by a user? Is BOINC allowed to use all of the CPUs at 100%? Is BOINC allowed to use enough memory that it can keep the CPUs busy the entire time BOINC is allowed to run? ...so some oddities are to be expected.

No fault of boincstats, it's just simply not easily extrapolated from the stats data provided by the standard BOINC server code.

The Average credit per CPU second column should rank CPUs but doesn't look right...

---

What Modsense said is true, that chart averages out all processors attatched to the project, even from people who no longer crunch, or only crunch for a little bit in the day. That is why my processor is so far down on the list, around 200 last time I checked, and my personal average is quite good with a mild overclock.
But...that 905e is high on the list with very few users and this is not a processor that is overclockable. This is not a processor you buy to overclock, it's an undervolted model meant to save on power.

---