Monday, October 13th 2008

Core i7 940 Review Shows SMT and Tri-Channel Memory Let-down

As the computer enthusiast community gears up for Nehalem November, with reports suggesting a series of product launches for both Intel's Core i7 processors and compatible motherboards, Industry observer PC Online.cn have already published an in-depth review of the Core i7 940 2.93 GHz processor. The processor is based on the Bloomfield core, and essentially the Nehalem architecture that has been making news for over an year now. PC Online went right to the heart of the matter, evaluating the 192-bit wide (tri-channel) memory interface, and the advantage of HyperThreading on four physical cores. In the tests, the 2.93 GHz Bloomfield chip was pitted against a Core 2 Extreme QX9770 operating at both its reference speed of 3.20 GHz, and underclocked to 2.93 GHz, so a clock to clock comparison could be brought about.

The evaluation found that the performance increments tri-channel offers over dual-channel memory, in real world applications and games, are just about insignificant. Super Pi Mod 1.4 shows only a fractional lead for tri-channel over dual-channel, and the trend continued with Everest Memory Benchmark. On the brighter side, the integrated memory controller does offer improvements over the previous generation setup, with the northbridge handling memory. Even in games such as Call of Duty 4 and Crysis, tri-channel memory did not shine.
As for the other architectural change, simultaneous multi-threading, that makes its comeback on the desktop scene with the Bloomfield processors offering as many as eight available logical processors for the operating system to talk to, it appears to be a mixed bag, in terms of performance. The architecture did provide massive boosts in WinRAR and Cinebench tests Across tests, enabling SMT brought in performance increments of roughly 10~20% with general benchmarks that included Cinebench, WinRAR, TMPGEnc, and Fritz Chess. With 3DMark Vantage, SMT provided a very significant boost to the scores, with about 25% increments. It didn't do the same, to current generation games such as Call of Duty 4, World in Conflict and Company of Heroes. What's more, the games didn't seem to benefit from Bloomfield in the first place. The QX9770 underclocked at 2.93 GHz, outperformed i7 940, both with and without SMT, in some games.

Source: PC Online
Add your own comment

91 Comments on Core i7 940 Review Shows SMT and Tri-Channel Memory Let-down

#1
PP Mguire
Holy Jesus you guys know to much! Yall sound like my damn dad where this signal thing is concernd.
Posted on Reply
#2
niko084
PP Mguire said:
Holy Jesus you guys know to much! Yall sound like my damn dad where this signal thing is concernd.
Some of us are into more than computers... ;)
Posted on Reply
#3
PP Mguire
Yea he has a few electronic and other degrees besides audio engineering. When he starts talking about that kinda stuff i get so lost. :(

But yea what they where saying, its like the old days of dual channel. You dont need a special kit because its all the same. I never ran matching ram and it ran in dual channel.
Posted on Reply
#4
niko084
masterbw2000 said:
Please prove it to us later. But what's the point if we already know the answer? :)
Sure, you will get like some points more for trying to enable tri-channel on dual-channel type of RAMs thus making tri-channel to look bad.
But the real tri-channel will give you MORE points, that's the fun part.

Will you be getting the evaluation samples? or the mass prod. version?
So now its Fake VS Real Tri-Channel huh....

Buddy take a hike or show some evidence outside of marketing hype.
Posted on Reply
#5
BOSE
masterbw2000 said:


2. Certified version of Tri-Channel memory shows large performance gain.
3. Circuitry differences.
Says Who?? Where is it stayed?
Posted on Reply
#6
Morgoth
can you tell me what pysical and technical stuff is differend from a dual pack/ single modules from tri channel pack

i'm not talking abouth the package

i pre-ordered the msi x58 epclipse board and the webshop allready sended it to me ;)
Posted on Reply
#7
masterbw2000
That's definitely a secret as of now from the memory manufacturers but they did say that there are differences. Like I said earlier, I doubted initially too and tried 3 identical sticks out of two dual chan kits, compared with just 2 sticks, then benchmarked it.
Posted on Reply
#8
Morgoth
and how many % increas was it? got a source?
what memmory modules did you used?
Posted on Reply
#9
niko084
Morgoth said:
and how many % increas was it? got a source?
what memmory modules did you used?
And how about a few screen shots of the benches...
Posted on Reply
#10
masterbw2000
Sorry, I would love to but unfortunately I can't, and I am sure ya'll understand.
Please just wait till Morgoth gets his stuff and then we will find out. Thank you.

Patience is a virtue.
Posted on Reply
#11
BOSE
masterbw2000 said:
Sorry, I would love to but unfortunately I can't, and I am sure ya'll understand.
Please just wait till Morgoth gets his stuff and then we will find out. Thank you.
Morgoth is allowed to post the info and screen shots and you cant?? Thats BS dude.

You sir, are a liar and hypocrite.
Posted on Reply
#13
niko084
Mussels said:
Tri channel will help prevent bottlenecks as we all go 8GB+ in our systems, and i think we need benchmarks with MORE than a measly 2GB of ram, before calling it a failure.
I think this is more the issue, than the idea that its not running in tri channel.
Posted on Reply
#14
BOSE
masterbw2000 said:
Dear BOSS, thanks for the comment.

I don't need to please you.

http://forum.coolaler.com/showthread.php?p=2185937
Look at the ram sticks in this picture two of them are they same and one of them is different. Durrrrr!!!! :roll:

http://www.coolaler.com.tw/coolalercbb/INTEL_I7_EXTREME965_X58/30.JPG


p.s. He is using G.Skill RAM, and they haven't made Tri-Chanel kits yet!

A better pic of his RAM. Two black sticks and one silver stick.

http://www.coolaler.com.tw/coolalercbb/ECS_9800GTX+SLI/12.JPG
Posted on Reply
#15
swaaye
This CPU isn't designed for games that can only use one or maybe two threads. Nehalem is probably going to look disappointing if you want more speed for such apps. But man, if you give this thing an app that can pound all 8 threads, it eats every other CPU alive. The whole architecture is designed to extract more performance from each core when there are many threads.

I don't know of any game that will really leverage the design here. SupCom is probably the most aggressively parallelized game engine and it can't peg more than about 3 cores. This chip may be ahead of its time...
Posted on Reply
#16
PP Mguire
That dude just got owned by his own linkage. Take a hike bro these guys dont tolerate liars!!
Posted on Reply
#17
masterbw2000
I knew you'd say that.
So take a look at the results from certified ones later.
Posted on Reply
#18
DarkMatter
masterbw2000 said:
Dear BOSS, thanks for the comment.

I don't need to please you.

http://forum.coolaler.com/showthread.php?p=2185937
http://forum.coolaler.com/showthread.php?t=191847
In those links (in the second one, with 6 identical sticks) we can clearly see how it's not benefiting from the triple channel even one bit. The memory running at 1800+ Mhz is ~40% faster than 1333 Mhz. If we add that 40% to the ~15000 MB/s results in the OP graphs we get no more no less than a theoretical ~21000 MB/s, around what the guy is getting with his overclocked PC. No benefit with triple channel.

masterbw2000 said:
I knew you'd say that.
So take a look at the results from certified ones later.
So where are those results? Or are the ones in the second link (which you added after an edit) the ones with certified modules? :roll:
Posted on Reply
#19
BOSE
This is with new A-Data Tri-Chanel Kit. Its posted on this website btw.

link - http://www.techpowerup.com/73397/A-DATA_and_ASUS_Demonstrate_Intel_Nehalem_s_DDR3_Performance.html

pic - http://www.techpowerup.com/img/08-10-09/a-data-x58-ddr3-bench.jpg At 2000Mhz




And this is same test done by the guy from the link you provided earlier. He is using G-Skill RAM.

http://www.coolaler.com.tw/toppc/I7920/20087.JPG At 1867Mhz


As you see, the read and write speeds of both RAM kits is almost the same. Thus proves the point that you just got OWNED!
Posted on Reply
#20
niko084
Alright chill out guys, all we are really showing is that with "current" software that "has" been used to test tri channel ram, we see nothing that makes us want to buy it.

While I doubt you will need a tri-channel kit for it to work, I do believe that tri-channel will show its benefits down line with proper application.
Posted on Reply
#21
DarkMatter
BOSE said:
This is with new A-Data Tri-Chanel Kit. Its posted on this website btw.

link - http://www.techpowerup.com/73397/A-DATA_and_ASUS_Demonstrate_Intel_Nehalem_s_DDR3_Performance.html

pic - http://www.techpowerup.com/img/08-10-09/a-data-x58-ddr3-bench.jpg At 2000Mhz




And this is same test done by the guy from the link you provided earlier. He is using G-Skill RAM.

http://www.coolaler.com.tw/toppc/I7920/20087.JPG At 1867Mhz


As you see, the read and write speeds of both RAM kits is almost the same. Thus proves the point that you just got OWNED!
Hmm it's interesting that 200 Mhz more didn't help increasing the bandwidth over what 1800 Mhz offered. Considering 20000 MB/s is also below what DDR3 1800 Mhz should offer in comparison to 1333 Mhz's 15000 MB/s (considering same efficiency in both cases), I think we can take one note: Nehalem CPU when overclocked to ts limits, in that specific benchmark can only offer 20000 MB/s and probably on stock it can "only" give 15000 MB/s, which is much better than Core2 anyway. My point is that maybe, in that benchmark, the memory is bottlenecked by the CPU!

It's not surprise to me anyway: Memory performance has increased a lot in recent years, while CPU doesn't. We have gone from memory bottleneck situations before 2004 to diminishing returns with more than DDR2 667 in 2008.
A CPU that is 33%-50% faster than Core2 will never be able to use a lot more than 33-50% more memory bandwidth than Core2, but due to proper usage of DDR3 and the IMC, the available memory bandwidth is much much more. Also thanks to the IMC it CAN couple with more REAL bandwidth that that 33-50% increase, but it still has a limit.

Adding a third channel increases the theoretical bandwidth but because the CPU can't work with such high bandwidth we see no gains. Keep in mind DDR3 1333 has a theoretical peak bandwidth of 10800 MB/s per channel (21600 MB/s in dual, 32400 in triple channel) so at 15000 MB/s in DC is way below that mark. On the other hand in the charts we can see the single channel one is sitting at 10300 MB/s, pretty close to it's peak bandwidth, one more detail that makes me think I'm right about this.
Posted on Reply
#22
Basard
summing it all up... sorta.... ya, this cpu will be cool, when crysis 6 comes out... lol
Posted on Reply
#23
PP Mguire
Im glad im just looking at an e8400. These seem to be a whop.
Posted on Reply
#24
DarkMatter
Well, I wouldn't say so. Just because Nehalem doesn't seem capable of using all the bandwidth it has unlocked by himself, that doesn't mean it's a fail. It's still way faster than Core2 clock for clock and has a lot more bandwidth even on single channel mode! SMT also seems to work very well.

It's maybe not worth it for games, but for everything else is faster enough to justify the expenses for many people (not me TBH), just as any other new CPU. Don't forget it's suposed to be aimed at the server market. Off course an upgrade from a C2Q is not worth it either, but if you were to build a completely new PC and you care about more than gaming, Nehalem is worth a look or two.
Posted on Reply
#25
FordGT90Concept
"I go fast!1!11!1!"
I think what it really boils down to is the IMC in Nehalem is over engineered and slower because of it. The designed the IMC to handle not just four cores, but eight, and maybe even more (as much as 72). I doubt we will see any significant changes to the IMC as far off as Sandybridge chips.

But yeah... maybe there is a bug in it. When AMD shifted to IMC, they got a huge memory performance boost. Intel appears to be taking a hit instead. CIS is a more complex scheme but you'd think that would show throw in more than just gaming benchmarks.

I just wonder how what Nehalem runs with FB-DIMMs.
Posted on Reply
Add your own comment