LGA 1156 Core i7s & Core i5s Reviews

[Clairvoyant129]

Make that 2Ghz

Back in reality,on average the Bloomfield is some 20-25% faster per clock than Deneb. You can make the Mhz numbers for "needed" Deneb from Bloomfied's clocks. The thing is that Bloomfield has a higher than that (average) lead in some workloads while in other it's not so much faster or is equal or even slower in few. So what you get in the end is about 20-25%,which is a healthy lead,a similar lead over Penryn which is some 5% faster than Deneb,which in turn coincides with a rough 20% lead every new intel design brought over previous one since P4 was sent to retirement(P4 not counted as "previous").

20-25%? Since when was the i7 920/i5 750 slower than the PII 965? I'm guessing you're not counting turbo boost since it's a "cheat" huh?

Penryn only 5%? Since you love quoting Anand, read this review and get back to me.

http://www.anandtech.com/cpuchipsets...spx?i=3551&p=8

Unlike you guys (v_rr, gosh and including yourself), I don't have an ulterior motive.

[gosh]

Thats right, that is why there are over 30 graphs. So you CAN get a firm understanding of each chip's benefits and pitfalls and see whether certain benchmarks favor what you (the end user) is going to be using their system for.

Benchmark apps are for a reason...

Today 99% can buy any processor about X3 or above and they will not notice any difference in speed.

[Hornet331]

Make that 2Ghz

Back in reality,on average the Bloomfield is some 20-25% faster per clock than Deneb. You can make the Mhz numbers for "needed" Deneb from Bloomfied's clocks. The thing is that Bloomfield has a higher than that (average) lead in some workloads while in other it's not so much faster or is equal or even slower in few. So what you get in the end is about 20-25%,which is a healthy lead,a similar lead over Penryn which is some 5% faster than Deneb,which in turn coincides with a rough 20% lead every new intel design brought over previous one since P4 was sent to retirement(P4 not counted as "previous").

ehh... you know with 20-25% faster you said the same thing as Clairvoyant... maybe a differenz of 100mhz...

[informal]

ehh... you know with 20-25% faster you said the same thing as Clairvoyant... maybe a differenz of 100mhz...

3.6Ghz deneb should overall match 3Ghz bloomfield(without turbo).

[villa1n]

Lynnfield PCI-Express Gaming Performance
http://www.bit-tech.net/hardware/cpu...-performance/1

why such a low res in this test... it cant hope to show gpu bottlenecking at least 1080p for xfire/sli setups...

[onethreehill]

The Lynnfield Followup: Turbo Mode and Overclocking Investigated
http://anandtech.com/cpuchipsets/showdoc.aspx?i=3640

[Xoulz]

Today 99% can buy any processor about X3 or above and they will not notice any difference in speed.

Butt test and benchmarks are two different things bro... we need deltas, so as a community, we can discuss the finer points of each.

[gosh]

Butt test and benchmarks are two different things bro... we need deltas, so as a community, we can discuss the finer points of each.

But it would be interesting if it was possible to discuss the technology behind

[Clairvoyant129]

3.6Ghz deneb should overall match 3Ghz bloomfield(without turbo).

Not even close,

http://www.pcgameshardware.com/aid,6...iewed/Reviews/

Now back to reality.

[informal]

Not even close,

http://www.pcgameshardware.com/aid,6...iewed/Reviews/

Now back to reality.

Try again:
Plethora of tests,965BE 6% behind 2.93Ghz i7 with SMT and turbo: http://www.computerbase.de/artikel/h...formancerating

Also a lot of tests and comparison between 3Ghz CPUs(Deneb,Yorkfield,Bloomfield): http://www.planet3dnow.de/vbulletin/...#content_start

[Chumbucket843]

Try again:
Plethora of tests,965BE 6% behind 2.93Ghz i7 with SMT and turbo: http://www.computerbase.de/artikel/h...formancerating

Also a lot of tests and comparison between 3Ghz CPUs(Deneb,Yorkfield,Bloomfield): http://www.planet3dnow.de/vbulletin/...#content_start

http://www.bit-tech.net/hardware/200...ition-review/4

[informal]

http://www.bit-tech.net/hardware/200...ition-review/4

We are talking about an average performance rating between these CPUs across various types of workloads,not about handpicking the workloads that i7 excels at(3d rendering and video encoding). Nice try though .

[gosh]

Now back to reality.

http://www.phoronix.com/scan.php?pag...ynnfield&num=1
http://ixbtlabs.com/articles3/cpu/in...7-lga1156.html

[Olivon]

http://www.phoronix.com/scan.php?pag...ynnfield&num=1

Update: after starting to see a flow of Windows-based reviews today, it looks like there are some more serious Linux + Lynnfield problems at hand, which we are currently investigating.

AMD is a little bit late. Hope Thuban and new revision will come soon.

[gosh]

AMD is a little bit late. Hope Thuban and new revision will come soon.

it is the opposite. Intel now has one processor that is equal to phenom II 955. AMD hadn't any real competition in that area before i5

[Xoulz]

But it would be interesting if it was possible to discuss the technology behind

That is what we are doing, isn't it?

Coincidentally, many are using graphs to illustrate certain points... that somehow you are saying is invalid, because YOU have different results on your butt dyno.

I find that rather odd.

[gosh]

That is what we are doing, isn't it?

Where?

How do cache design from phenom and bloomfield differ and how does that effect performance in different loads?
How does prefetchers on intel work and how does prefetchers effect performance in different loads?

[villa1n]

http://www.bit-tech.net/hardware/200...ition-review/4

http://www.bit-tech.net/hardware/200...ition-review/5

I d be suspect of a review that shows the 955 at stock having better numbers than a 965 at stock :/ , I m not really sure what the arguement is here, skimming these past pages... we all know Deneb and Penryn are very close within 5-10% depending on the bench/application/game, and we all know that bloomfield has a much higher ipc than both... So i m not sure where this discussion is going? Are we talking about threshold performance, where we need things to only be a certain speed within a cost window, ie i have 1400$ to spend and i need these tasks to be done, which is the best platform to achieve? Or are we comparing absolute values in which case i m not sure why the discussion isn't bloomfield vs LynnField and more importantly, i wish there was more testing of the pci-e bandwidth contraints compared, ie 4870x2's or gtx295's sli'd / xfired as with the 5870 and eventually the gt300 coming out, and both.. single versions of those cards being more powerful than both the 4870x2/gtx295.. seems pretty relevant that people would want to see how that compares, as the one review i did find that shows the lynnfield vs bloom on xfired 4870x2's shows what seems to be a choke in min FPS and in AVG fps, any links to more of this testing would be appreciated, because I am seriously considering one of these platforms for my next build,...
I think the dual vs triple channel and has been maxed out .. we all know heavy memory bandwidth apps are going to need bloom or more memory, and there is a plethora of tests showing the impact and latency advantages/disadvantages..

[Computurd]

Yes i am very curious of the ramifications of the pci-e slots in the future, as well. EVGA hosted a p55 discussion panel with their designers two days ago, and in it they seemed to be intentionally vague, and even said things like "with current gen"...sounds to me like they are waffling on the impact, because they really arent sure.

[Chickenfeed]

I'm assuming that is in regards to SLI / CF but that vagueness from EVGA isn't a good sign. All this marketing with P55 for doing multigpu and if it turns out to be a poor show with next gen gpus that going to be a kick in the pants for board makers as well as early adopters.

[Computurd]

yes they even admit to the fact that you arent gonna get high vga clocks on the boards equipped with the nf-200...which amazes me...what i dont get is this,
When was the last high powered GPU released that was x8 out of the box??? this must be a great trivia question....
so the REAL question is:
if its been so long since x8 was standard, why in the world are we manufacturing x8 SLI boards and slapping big fat CLASSIFIED and ENTHUSIAST and OVERCLOCKING stickers on them?: on this chipset designed for OEM!!!
almost comical
its like they are going backwards!!!

[Clairvoyant129]

Try again:
Plethora of tests,965BE 6% behind 2.93Ghz i7 with SMT and turbo: http://www.computerbase.de/artikel/h...formancerating

Also a lot of tests and comparison between 3Ghz CPUs(Deneb,Yorkfield,Bloomfield): http://www.planet3dnow.de/vbulletin/...#content_start

How cute,
You see, with those reviews I don't even know what kind of a test platform they used. It's funny how you need to go find obscure reviews in a foreign language.

You love quoting Anand but when I used his results to backup my conclusions, you go find some reviews in a different language and god knows what kind of a OS they used (since Linux favors PIIs and you seem to pick reviews where they used Linux as main OS)... thanks but it's obvious you can't pick any major reviews since they contradict your conclusion.

And lol, don't even bother with Planet3D... must be your favorite site huh?

[villa1n]

I'm assuming that is in regards to SLI / CF but that vagueness from EVGA isn't a good sign. All this marketing with P55 for doing multigpu and if it turns out to be a poor show with next gen gpus that going to be a kick in the pants for board makers as well as early adopters.

Exactly.. Well I guess we ll see when the nda lifts for the 5870 how those do in xfire and if they bottleneck x8 pcie2.0 on lynnfield.

[v0dka]

How cute,
You see, with those reviews I don't even know what kind of a test platform they used. It's funny how you need to go find obscure reviews in a foreign language.

Yeah because only English speaking reviewers know they are doing, what do Germans know about computers. Silly little country with only 80 million people.

[gosh]

You love quoting Anand but when I used his results to backup my conclusions, you go find some reviews in a different language and god knows what kind of a OS they used (since Linux favors PIIs and you seem to pick reviews where they used Linux as main OS)... thanks but it's obvious you can't pick any major reviews since they contradict your conclusion.

don't you think it is strange that most sites are testing same apps?
if tests are perfect that you should get suspicios because testing different apps is not perfect.
What I think is the best reviews is those that show up after those done on the release date. Reviews done in other languages often test other applications.