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

Xeon 45nm quad X3350 (same as Q9450) in stock

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
My thinking exactly. That is why I went ahead with the Q6600 last week. I just couldn't see it giving $130 worth of extra performance.

If they are both clocked the same, performance will be almost identical. Only difference will be the cooler temps on the 45nm. Obviously, that leaves more headroom to OC. Id imagine that it is extremely rare if a q6600 hits 4.0 on air. I have a feeling the 45nm chip will be doing that regularly. Maybe it wont be enough to justify the extra $130, but its pc gaming. An enthusiast will pay to play. Even if it means the extra $130 for only 5-10% performance increase.
 
I'm fairly interested in the PC article that stated the Xeon had different pre-fetchers. I've always been under the impression that the Xeon was merely a workstation class CPU. In the past it's often had more cache, but of course low end version of the Xeon and high end desktop have been identical except for the name (and SMP capable).
Xeons are also SMP capable, meaning they can run in multi-socket boards (but beware, this particle 775 Xeon is a single workstation CPU...because there are no dual socket 775 boards), you need to get socket 771 Xeons and obviously boards that run 771 to utilize these chips (skull trail for example)

Also, speculation has been Xeons in the past were binned as the better CPU's (much like AMD's opteron). So in theory, in the past, they've been better OC's or run cooler, or more stable at higher temps. But thats speculation and is it true today? who knows. Xeon's are meant to sell to large enterprise/server/workstation businesses.

So in closing, I think they're exactly the same chips except for a different name and possible SMP capabilities, or simple small pin type mods, nothing serious such as different optimized pre-fetchers or architectural changes. I could be wrong though, it would take a pretty credible source for me to believe otherwise.
 
These are approximate but very close

E8200 = $167
E8400 = $189
E8500 = $316
Q9300 = $266
Q9450 = $316
Q9550 = $514
 
firstly, I hope I'm wrong as well, I was just mentioning something I read. The magazine is a british one called Custom PC, in issue 53, they had a large cpu test, where they tested pretty much every intel and AMD chip available, they stated in the section about the X3200 series xeons that they were 3% faster in cinebench and folding@home, but 25% slower in games, than an identically clocked Q6600.

there is a transcript of that review here if you want to read it. They do a whole suite of benchmarking, the results are in the magazine, not sure if they're on the website.

I dont expect that ALL Xeons are like this, but if Intel tweaks the prefetchers on some models to optimise them for their intended use, why not all?

I just thought perhaps I should mention it in case hardcore gamers go out and buy one, only to be disappointed with the results!

the issue is also mentioned here on the website in the review of the skulltrail platform if you're interested
 
If they are both clocked the same, performance will be almost identical. Only difference will be the cooler temps on the 45nm. Obviously, that leaves more headroom to OC. Id imagine that it is extremely rare if a q6600 hits 4.0 on air. I have a feeling the 45nm chip will be doing that regularly. Maybe it wont be enough to justify the extra $130, but its pc gaming. An enthusiast will pay to play. Even if it means the extra $130 for only 5-10% performance increase.
I really doubt we will be seeing high clocks out of the 45nm quads. The low FSB wall and low multi don't make for a good combination. Unless for some crazy reason these retail chips blow the QX FSB out of the water, which I don't see happening. I'm guessing people are going to be stuck around 450FSB.
 
There are a lot of members here who have dual and quad core Xeons that never have mentioned anything about a 25% loss in gaming performance.

I highly doubt there is a difference between the xeons and Core 2 Series.
 
I really doubt we will be seeing high clocks out of the 45nm quads. The low FSB wall and low multi don't make for a good combination. Unless for some crazy reason these retail chips blow the QX FSB out of the water, which I don't see happening. I'm guessing people are going to be stuck around 450FSB.

qft. as stated, if i dont move to the QX 45nm quad - i'll be waiting for the q6700 with it's 10x multi in april.
 
What bothers me is that they take orders with the false pretense that they have the processors in stock and then turn around and send the email telling people that they are out of them. That is unfair and fraudulent business practices. They get caught doing this practice in a lawsuit they will get hammered.

It isnt fraud at all, their inventory or website system probably couldnt keep up with the possible hundreds of orders rolling in at once, who is to say someone esle didnt place their order literally milliseconds before you, they got the last one, then system said OH crap, we have no more....



So sue them since you seem SO sure it is illegal and fraud... but first you may want to read over all of their fine print on their website, as like most stores, i am sure the covered their butt's over things like this.


I am sure a quick email back telling them to cancel the order, and they will, now if they refuse to cancel your order and refund the money - then you have a lawsuit,
 
... the X3200 series xeons that they were 3% faster in cinebench and folding@home, but 25% slower in games, than an identically clocked Q6600.....

I just thought perhaps I should mention it in case hardcore gamers go out and buy one, only to be disappointed with the results! ...
I really wonder what sorts of games these guys were testing. I've monitored Crysis several times and found my four Q6600 cores only about half busy, while my overclocked 8800GTX is pegged. The GPU is clearly the bottleneck - not the CPU.
 
the games used are FEAR, STALKER and need for speed:carbon, I believe they use supreme commander with some setups, but it's not an official game used.

those games are the 3D graphics benchmarks, I believe the Crysis demo and supreme commander were used in the CPU test, all the LGA775 chips were tested on an asus p5k premium p5 board

you can download their benchmark media suite from www.custompc.co.uk if you want to test it.

apparently there is a CPU benchmark with the crysis demo where the player is running around blowing buildings up, the CPU is stressed trying to calculate all the debris effects, the test was run at 1280x1024 with all detail settings on medium, except world physics which was set to very high.

The supreme commander benchmark is the one built into the game at the same res as crysis, with 2xAA.

The results showed a 25% lower performance from an identically clocked Xeon.... make from that what you will.
 
Well, let's make sure we're talking about the same things here:

They were testing the Xeon 3200 series; this thread is talking about the Xeon 3300 series. And I'm also not entirely convinced that the Xeons really have a reason to be tuned differently -- especially when it results in such huge performance disparity on certain apps.

Another complaint: that article didn't mention video hardware, drivers, operating system, if a 32-bit or 64-bit version of OS and/or application were used, application versions and patch levels, memory timings, among a HUGE pile of other items that are VERY pertinent to performance discussions.

Basically, there's enough wrong with how they documented this review that I don't have any compelling reason to believe it.
 
Well, mine just arrived:

FPO/BATCH#: L803B196
S-SPEC: SLAX2

(Man, the Dobbsites must be loving this)
 
Well dammit, stop posting and start overclocking!!1!1 What forum do you think this is, the Stock-clockers-chitchat forum? :beer: ;)

Unfortuantely, I'm at work. I won't be able to pop it in until after I get home (about 5-6 hours from now). I will definately post results as I get them.
 
Naw, was just ribbing you good since I'm completely jealous. Definitely want to see how these clock compared to your Q6600... I really want to see one paired with an X38 chipset.
 
Well, let's make sure we're talking about the same things here:

They were testing the Xeon 3200 series; this thread is talking about the Xeon 3300 series. And I'm also not entirely convinced that the Xeons really have a reason to be tuned differently -- especially when it results in such huge performance disparity on certain apps.

Another complaint: that article didn't mention video hardware, drivers, operating system, if a 32-bit or 64-bit version of OS and/or application were used, application versions and patch levels, memory timings, among a HUGE pile of other items that are VERY pertinent to performance discussions.

Basically, there's enough wrong with how they documented this review that I don't have any compelling reason to believe it.

firstly, I'm aware that we are talking about a different series of Xeons, I was just stating that if intel have done it before, they might do it again.

secondly, they did provide detailed specs on almost all of the equipment used, they probably didnt note all of that down on the website as they need to sell magazines, so are obviously not going to provide everything for free.

From the magazine:

The OS used was vista 32-bit edition, running from a Samsung Spinpoint P120S hard disk, the graphics card was a 320MB BFG Geforce 8800GTS. The system memory was 2GB of corsair XMS2-8500. It stated that all the latest graphics and chipset drivers were used. Presumably they were the latest versions as of January 2008. They would have no reason to use different memory timings, although if they did, it wasnt mentioned.

I'm simply the messenger here, I dont want to believe it as I've been considering a Xeon as my next chip, but I just wanted to mention that there could be a difference. Why would intel release two identical chips, but brand them differently? to my mind, it makes little sense.

Anyway, I'm done talking about it, good luck overclocking and I hope you all have great experiences with whatever choice you make.
 
Kyussinchains...
Thanks for the links to those interesting articles. :) I have the x3220 and always wondered if indeed it was the same as the q6600. I will add my 2 cents and say that there has to be some difference between them. This is the question I sent to Gigabyte asking them why it was NOT on their CPU Support List (no xeons are for my mboard) and their reply...

No problems, seems to be working, but wondered if the X3220 processor is OK to use with this board? It is not on the "CPU Support" list. Do you plan to include it, or evaluate it, or should it NOT be used? I picked the XEON as it gives a bit better temp specs than the Quad desktop processors.

hello,
sorry , X3220 is not on the support list it has different vcore voltage that can not run on this board .

Sooooo... not a good answer, but it does work OK. The CPUID comes up in bios as 06FB for my SLACT processor which I believe is the same as the Q6600 GO SLACR. Maybe BIOS needs to be tuned for the XEON, but if both give the same CPUID, the board does not know which it is?? Another question is what boards do support 775 socket XEON processor? I don't think any 775 gigabyte boards supposedly do. So, are they identical? I doubt it. :eek: Is there a performance hit in games? dunno, I don't game.
 
Back