Low End Xeons.

ATIIsBetter

[H]ard|Gawd
Joined
Feb 19, 2005
Messages
1,202
I was checking the prices on the Core 2s today when, "OMG XEON FOR $200 LOL!"

So.... I was reading the specs and it looks like I can basically get an E6300 with 4M cache for $20 more. I feel like there's some catch though. Even though the Xeon 3040 is an E6300 with more cache(As if the higher Core 2's don't have 4M), will it not run on the first mid-range $130 Core 2 baord I randomly pick? Any thoughts would be nice.

The Xeon 3040:
http://www.newegg.com/Product/Product.asp?Item=N82E16819117102
 
It should work... the Xeon 3000 series ARE conroe cpus. I have yet to try one, so i can't give you a 100% yes or no. The reason for the new nameing scheme is becuase of the new single cpu server chipsets that have come out... and you can't sell a Server board without a Xeon. :D
 
I believe they are to be used in the 3000/3010 chipset boards, but that may have changed...
 
ND40oz said:
I believe they are to be used in the 3000/3010 chipset boards, but that may have changed...


That's the intention... but they are LGA775 CPUs and validation on those chipsets was done with Conroe CPUs as well as the P4 family. It's a desktop CPU, labled as a Xeon. Really though... there are no differences between ANY of the xeon family and the desktop family anymore.. with the exception of the package. It should work.
 
I hope that the Op will let us know if he can get it to run on a desktop board :)
 
Hmm... my documentation says that those are 2mb cache parts.

Code:
Server	Intel® Xeon®	Conroe	3040	HH	80557KH0362M	1.86GHz	1066	2M	FC-LGA6	Full	B-2	SSPEC	SL9TW	885516	Prod	Sep '06 launch		
Server	Intel® Xeon®	Conroe	3050	HH	80557KH0462M	2.13GHz	1066	2M	FC-LGA6	Full	B-2	SSPEC	SL9TY	885515	Prod	Sep '06 launch		
Server	Intel® Xeon®	Conroe	3060	HH	80557KH0564M	2.4GHz	1066	4M	FC-LGA6	Full	B-2	SSPEC	SL9TZ	885514	Prod	Sep '06 launch		
Server	Intel® Xeon®	Conroe	3070	HH	80557KH0674M	2.66GHz	1066	4M	FC-LGA6	Full	B-2	SSPEC	SL9U2	885513	Prod	Sep '06 launch

methinks newegg mislabeled something.

Edit... check on Intel.com. They don't list that CPU as being 4mb either. Newegg mislabled.
 
All the documentation I've seen lists the 3040 and 3050 as 2 mb cache as well.
 
Poncho said:
Hmm... my documentation says that those are 2mb cache parts.

Edit... check on Intel.com. They don't list that CPU as being 4mb either. Newegg mislabled.
I wish Intel.com was usable, but they check a friggin cookie on every page, constantly sending me back to the stinking intro :mad:
 
Apparently these might actually be 4 mb cache, atleast that's the discussion over at AT. Poncho, anyway you can confirm or deny that intel is going to update the specs and that these are actually 4 mb cache chips?
 
ok...there still seems to be some confusion as to if the Xeon 3050 has 4mb L2 cache or not
http://www.newegg.com/Product/Product.asp?item=N82E16819117103
(as newegg still is showing that it does have 4mb L2 at the time of this posting)

what has happened is that Intel had a misprint on their website that showed the two lower-end socket 775 3040 and 3050 Intel Xeon Conroe CPUs as having a 4mb L2 cache the same as the 3060 and 3070
(as seen captured here > http://www.tipidpc.com/viewtopic.php?tid=74314)
...and this is where newegg got their information from....this has now been corrected on the Intel website, and the 3040 and 3050 Xeon CPUs now show that they really do only have a 2mb shared L2 cache
...I'm guessing that further confusion was caused when people contacted Intel to ask about this, and either the Intel representatives used this same misprinted information, or they gave information about the socket 771 Woodcrest Intel 5000 series Xeon CPUs

and to show proof that the 3050 Xeon does indeed really have 2mb L2 cache,
here is a screenie of CPU-Z with a 3050 >
2mrz21y.jpg
<



now there are differences between the socket 775 Xeon 3000's and the normal Conroe/Allendales:
Previously mentioned, the desktop and server processors used identical instruction sets. To a degree this statement is incorrect.

Each processor does contain the following advanced processing level instruction sets:
SSE, SSE2, SSE3, SSSE3 (Supplemental), MMX, x86-64.

The grouping directive to comply with IA-32 is however slightly different.
x87 Floating Point Instruction - Same
SIMD Integer Instruction - Different
SIMD Floating Point Instructions - Different

These are the two main differences between the desktop and server variants of the Core Micro-architecture.

x87 Floating Point Instruction - A single instruction that improves x87 FPU floating point to integer conversions. This is identical to remaining Intel architectures i.e. Netburst.

SIMD Integer Instruction - One instruction that provides a specialised 128-bit unaligned data load.
With The Xeon based architecture the ability of this instruction is greater. Often more complex based instructions are interfaced with this ability for mathematical arithmetic and logic processing. These instructions can either be 32 or 64 bit designed.

SIMD Floating-Point Instructions - Three instructions that enhance LOAD/MOVE/DUPLICATE performance.
Two instructions that provide packed addition/subtraction.
Four instructions that provide horizontal addition/subtraction.

Again these are capable of interfacing with other advanced complex instructions. Including decoded x86-64 complex instructions.

I know this is far beyond most readers knowledge but still it does show you that a Xeon is capable of working with mathematical instructions much better. However as the Xeon is capable of doing this it is less capable of using SIMD level instruction sets for game interfacing as the instructions have an advanced mathematical base, not physics/geometry.

-"The_Manual"

[edit]change the CPU-Z screenie link to img tags[/edit]
 
This post is OLD NEWS. I saw this 1 week ago, and posted it here, and we all agreed that it is 2MB cache, and NEWEGG did mislabel it.
 
2MB L2 or not, I wish I could have got a 3040/3050 instead of the 5110 at work. They decided to order me a "nice" computer and went with a Precision 690 workstation with expensive memory and a mediocre 1.6GHz CPU. :p
 
pxc said:
2MB L2 or not, I wish I could have got a 3040/3050 instead of the 5110 at work. They decided to order me a "nice" computer and went with a Precision 690 workstation with expensive memory and a mediocre 1.6GHz CPU. :p

The 3000 series Xeons won't fit in the socket of your board. You'd have to stick with the 51xx Xeons.

30xx = LGA775
51xx = LGA771
 
Poncho said:
The 3000 series Xeons won't fit in the socket of your board. You'd have to stick with the 51xx Xeons.
Yeah, I know. And a 3000 series won't fit in the dual LGA771 board on the stupid Precision 690. :p I wish they would have asked me to choose a system and I would have gone with an Optiplex and a fast Core 2 for the same price.
---

Anyone know of an overclocking utility that works with the Precision 690? 1333MHz FSB would get that system to 2GHz.
 
pxc said:
Yeah, I know. And a 3000 series won't fit in the dual LGA771 board on the stupid Precision 690. :p I wish they would have asked me to choose a system and I would have gone with an Optiplex and a fast Core 2 for the same price.
---

Anyone know of an overclocking utility that works with the Precision 690? 1333MHz FSB would get that system to 2GHz.

I haven't found any besides the one over at 2cpu. A guy over there did a bsel mod to get his 1066 5120 to run at 1333 like a 5140. Who the heck speced your system for you and why would they go with 1066 woodcrests...
 
newls1 said:
This post is OLD NEWS. I saw this 1 week ago, and posted it here, and we all agreed that it is 2MB cache, and NEWEGG did mislabel it.

sorry newls1, but no one was able to confirm that information either way...and to an end to it all I was able to find a screen shot of a 3050...old news or not, it was all just people posting their 'e'pinions on web forums, and hopefully my post was able to clear that up

T Rush said:
ok...there still seems to be some confusion as to if the Xeon 3050 has 4mb L2 cache or not
http://www.newegg.com/Product/Product.asp?item=N82E16819117103
(as newegg still is showing that it does have 4mb L2 at the time of this posting)


and to show proof that the 3050 Xeon does indeed really have 2mb L2 cache,
here is a screenie of CPU-Z with a 3050 > http://i13.tinypic.com/2mrz21y.jpg <

newls1, if you did have this information(and a way to prove it) you really should have posted it here...you say you did, but this thread is what Google turns up for information about these CPUs on [H], and I don't see your post except for the one I quoted


I hope the information I posted as able to provide new knowledge to the [H]ard|Forum readers, and I think that this is also an important take on these CPUs:
even if these Xeon 3040 and 3050 CPUs are able to overclock to higher speeds than the E6300 and E6400, the gaming performance might be lower than them...at least thats what I gather from that quote...something along the lines of:

'the Xeon CPUs have better mathematical instructions as a trade off for poorer level instruction sets for game physics/geometry interfacing'
 
I talked to a guy who worked at Intel at a recent LAN party i went to. It was at the intel HQ in hudson, MA. And he told me the same thing.

T Rush said:
'the Xeon CPUs have better mathematical instructions as a trade off for poorer level instruction sets for game physics/geometry interfacing' [/b]
 
T Rush said:
:
even if these Xeon 3040 and 3050 CPUs are able to overclock to higher speeds than the E6300 and E6400, the gaming performance might be lower than them...at least thats what I gather from that quote...something along the lines of:

'the Xeon CPUs have better mathematical instructions as a trade off for poorer level instruction sets for game physics/geometry interfacing'

Sorry man... but that's simply not true. The 30xx series xeons are Conroe CPUs... there is no difference.
 
Back
Top