• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

And the Debate Continues

E

EvilAlchemist

2[H]4U
Joined
Jan 11, 2008
Messages
2,732
Anyone who has been around a while has seen a thread with "Xeon Vs Desktop" or something similar.

The thread usually ends two ways: 1) No Difference 2) Xeon are better for "Server Type Load" but ends there.

This started with the advent of Core 2 Duo Xeons in the LGA775 socket.

Many users found there was more availability and sometimes no price difference between the Xeon & Desktop Processor.

Since Intel does not publish any real details about what makes a Xeon different vs Desktop Processors, it mostly comes down to opservation and benchmarking.


Here is a quote from bit-tech.net article (taks about the Xeon 3200 Series)
(X3200 is the server equivlent to the Q6600)

However, the Xeon X3200-series CPUs aren't just Core 2 Quads with a different name engraved in the top of the heatspreader. As Xeons are aimed at servers and workstations, the four prefetchers are tuned for these sorts of applications, rather than standard desktop applications and games. As a result of this subtle tuning, the X3220 was 3 per cent slower than the otherwise identical Core 2 Quad Q6600 in our Media Benchmarks, and 25 per cent slower in games. However, in our Folding@home test and Cinebench R10, the X3220 was 3 per cent faster than the Q6600. Unless you're building a server or workstation, you're better off buying a Q6600 over an X3220.
Click to expand...

This seems to say that Xeons are better for High Load situations - perfect for Folding@Home.

Here is a quote from a XtremeSystems User

The Xeons are the best chips off the wafer in each bin, most would clock better, at lower temps, power and volts than the LGA-775 counterparts given the same platform, cooling and BIOS. They have far better stability and lower degradation over lifetime too and yup, they are specifically tuned for different tasks native to the server world. Since Intel doesn't make much public mention on this through documentation means we can't really know how and its full effects until rigorously and astutely benchmarked following a fixed methodology, OR, if a devoted review site carries this out for us. I doubt highly there is such a difference between the two as shown in Post#1 with low load applications such as desktop ones; it would matter and show up in server applications properly, esp. TPC-C, Linpack and the like, wherever you have very high core loads with high memory usage scenarios, the Xeon will show more performance optimizations. Maybe they added some latency in the cache algorithms to optimize for aggressive large array prefetching, that's all I'm thinking has been done.
Click to expand...

These comments & the above article would seem to indicate that the Xeon Processor Line is better for Folding@Home
The problem with this conclusion is it is directly opposite of what other users think.

So, I decided to run my own test since I am running Xeon & Desktop Systems to see if there was any difference.

Each system is the the Uni-Processor Client & WorkUnit 2493 (905 points)
Each CPU was loaded to 100% by running a F@H client on each core for 6 hours prior to get PPD numbers.
Each Client/ Core was locked to a single CPU core to elimate WU drift.
PPD value is from a single client and all cleints on CPU get the same PPD.
My live PPD stats can be view at http://www.evilalchemist.com if you wish to view all data.

  • Laptop -------> Desk T9300 (2.50 GHz 45nm) - 461.85 PPD
  • FileServer ---> Xeon X3350 (2.66 GHZ 45nm) - 502.52 PPD
  • WHS ----------> Xeon 3075 (2.66 Ghz 65nm) - 465.15 PPD
  • HTPC ---------> Desk Q6700 (2.66 Ghz 65nm) - 458.87 PPD


We know that the 45nm CPUs are faster clock for clock then the 65nm processors.

If you compare the Xeon 3075 vs Q6600 - there is a 6.28 PPD difference. Xeon Wins.

If you comare the Xeon 3075 vs T9300 - there is a 3.30 PPD difference. 65nm Xeon beats a 45nm Laptop CPU.
(There is a 160 mhz difference between the two CPUs - so performance would be about equal - cancels out the 45nm advantage)

That may not seem like much, but remember this is just one core of the CPU, a SMP unit would see even large % gains due to the multi-core nature of the unit.

It would seem based on these observations that the Xeons do perform better then their desktop counterparts.

One thing I can't control in these Tests is the Front Side Bus of Each CPU.
Xeon X3350 = 1333 MHZ
Xeon 3075 = 1333 MHZ
Desk Q6700 = 1066 MHZ
Desk T9300 = 0800 MHZ

I have read some reports that higher FSB does result in higher synthetic benchmark numbers but also shows no difference in real world tests.

I invite the Folding@Home users to review this, present your thoughts, and run your own tests to try and confirm my results.


 
As you eluded to in another conversation, even if they are the same clock for clock, the Xeon chips are going to give a better ppd/watt because of lower power consumption.

I don't have any Xeon chips to verify.
 
You did see differences, but I think those are within the margin for error, considering there are other variables at play:
-FSB differences (you mentioned this already)
-RAM speed
-How fresh the OS installation is
-What OS is installed
-Number of background processes
-Specific workunit and PRCG (even workunits with the same points value can show different PPD on the same CPU or GPU depending on the specific unit itself)
-Fahmon is also not the most accurate at calculating PPD

Not that I'm trying to discount your testing, but I believe that a lot more needs to be done before you can say anything conclusively. Also, specific testing would need to be done with SMP to see what the results are with that, since they might not even be in line with the uniprocessor numbers.
 
For me, hands down is the $$. If I can get a better deal on a Xeon for the same speed as the desktop processor, its the winnner. But if the desktop is the clear $$ favorite, well...

If you live next to a frys/microcenter and they are throwing 860's out the door with a motherboard for 249.99, well I would by that all day long. Unfornately, they wont box the X3440 (2.53 with hyperthreading) with a mobo and still be under 300.00. Otherwise, i'd buy those all day long.
 
-FSB differences (you mentioned this already)
-RAM speed - DDR2 800mhz Corsair in all systems
-How fresh the OS installation is - 1 month on each system
-What OS is installed - Vista on all but WHS (Xeon 3075)
-Number of background processes - Default Vista with desktop effects set to 0
-Specific workunit and PRCG (even workunits with the same points value can show different

On that note, each System showed the Same PPD and Frame Time

Example: Q6700 - all 4 client show the exact same PPD across the board with the same frame times on each unit.

That held true for each cpu. You can see on the my client status page how that holds true.

FAHmon may not be 100% accurate, I can post frame times from each WU to show they are 100% the exact same from each CPU.

I can never get each system 100% the same, but these do point to some very interesting variations.

I wish I could come up with a conclusion, cause I want the answer as much as you do!
 
Would a native notfred installation not be the best benchmarking tool for this? It eliminates some of the OS-related variables that Zero mentioned.
 
gwarren007 said:
For me, hands down is the $$. If I can get a better deal on a Xeon for the same speed as the desktop processor, its the winnner. But if the desktop is the clear $$ favorite, well...
Click to expand...

That is also a good point of debate. Xeons usually have a higher initial cost but if they get better PPD/watt .. it can work out to be a better deal in the long run.

My X5550 CPU's (no in this test) have a TDP of 95W. (My CPU actually users 87.5 watts at full load)
The i7 920 CPU have a TDP if 130W

So the Xeon will use less power then the i7 920 at the same speed

That is where the initial costs vs costs to run comes in

flogge said:
Would a native notfred installation not be the best benchmarking tool for this? It eliminates some of the OS-related variables that Zero mentioned.
Click to expand...

It would eliminate some of the variables involved, but I am unable to run that on some of my systems.
I can't take my HTPC (Q6700) & Laptop (T9300) down for that long to test.
I attempted to make each windows exactly the same to the best of my abilities.
I am hope some other users can test and post their results to see what data they get.
 
flogge said:
Would a native notfred installation not be the best benchmarking tool for this? It eliminates some of the OS-related variables that Zero mentioned.
Click to expand...

This is a very good point.

This brings back memories of the farms of headless boxen everywhere running the brand new SMP client.

Memories.
 
And who can forget the [H]ardoxen adoption agency. I got in on the tail end of that when the GPU2 client was just getting up and running.
 
lol....... good argument...

When it comes down to it for me, if I can get a Xeon thats the same specs internally, but is just a tad bit cheaper than its desktop rival, I will buy the Xeon.

My goal now is an i5.
 
FrEaKy said:
lol....... good argument...

When it comes down to it for me, if I can get a Xeon thats the same specs internally, but is just a tad bit cheaper than its desktop rival, I will buy the Xeon.

My goal now is an i5.
Click to expand...

getting an 860 i7 isn't all that much more.
 
Any idea how good or bad the Xeon (lynnfield) cores are overclocking compared to the desktop versions ?
 
If the X3440 pulls 3.4-3.8ghz, what is common for the i7-860? I am trying to decide between the two processors. I will be using a CM Hyper 212+.
 
Zero82z said:
Definitely aftermarket. The stock heatsink is only adequate for very light overclocking in any situation.
Click to expand...

Also be aware that xeon heatsinks are mounted differently then Core i7 heatsinks.

Make sure you use the correct one for your motherboard.
 
Slªde;1034950733 said:
If the X3440 pulls 3.4-3.8ghz, what is common for the i7-860? I am trying to decide between the two processors. I will be using a CM Hyper 212+.
Click to expand...

I'm running mine (i7 860) at 3.3 with default core voltage. I could certainly probably try and push it higher by adding Vcore, but it's nice and stable there with default. :)
 
You must log in or register to reply here.
Back
Top