Core i7-970 "cheapest" Intel six-core on shelves around the world

[Sam_oslo]

AMD offering is much better: value/$,value/watt

There are times I'm ashamed that I have Intel platform, like now when I hear people defending Intel monopolistic strateries and price policies.

About i7-970: 850$ funny joke. I'll stay 4c.

You are wrong, because value/$ is equal to value/value or $/$, and all of them are always equal to 1 , I don't know what you are trying to say with value/watt? but Intel's 32nm has better performance/watt. When it comes to superior (without competition) high-end CPU, GPU, etc .., the performance/droller is not good. Because the competition (or lack of it) influence the prices for high-end components.

I've already said this thingy got more expensive than expected. A superior high-end CPU, GPU, etc.. without a real competitor, has always a much higher price. It is not about Intel or defending their prices, it is just a fact and explanation about the role of competition.



It's not a shame to have a good Intel CPU, it is more shameful to accuse others. who are you referring to, and who is defending monopoly or high prices?

[PatRaceTin]

price is too high

[Nintendork]

AMD's best offering performs the same as 2 years old 920, and costs more too!

But that's the technology of the past, and totally uninteresting. 32nm is the norm now, and that's why this thingy got so expensive, there is no competition in high-end desktop CPUs from AMD yet.

i7 920/930 ~ X6 1055T
i7 950/965 ~ X6 1090T

More expensive? Really?


Maybe if we talk about superpi.

[EvilOne]

Another great chip!

Shame only a minority will be able to afoird one

*hugs Q6600*

[dartaz]

Should have said : " Core i7-970 "cheapest" intel six-core on shelves around the world"

We have six cores on desktop for under 200 bucks now. Oh and it can do 4Ghz on air too

But PII X6 @ 4GHz is not as good as i7 quad core at same clock in majority of cases



I wonder why people compare AMD 6 core to intel 6 core ?

[Splave]

It's called common sense.

You don't make a thread claiming $850 is the cheapest hex core in the world when you can buy FOUR of them for that price.

What you call 'AMD force' I call not being a thick eejit

Go buy 4 of them then? or by a 4core8thread 920

[gosh]

But PII X6 @ 4GHz is not as good as i7 quad core at same clock in majority of cases

Test multitasking performance (two or more applications) and applications where data isn’t using aligned (most of them).

[demonkevy666]

But PII X6 @ 4GHz is not as good as i7 quad core at same clock in majority of cases



I wonder why people compare AMD 6 core to intel 6 core ?

failed to show Nb speed
so bus speed bump on I7 930 has advantage there.

[xVeinx]

Test multitasking performance (two or more applications) and applications where data isn’t using aligned (most of them).

Even using linux, where the performance difference is minimal, there are noticeable differences in performance between the 1055T and the Core i7 920. Compilation time for software is still faster on the 920 (5-7%), and sequence alignment using indexing and a Wheeler-Burrows algorithm is still faster (about 8% or so, software is called Bowtie for those interested). The 1055T is no slouch for sure, but it isn't a complete 920 killer in every instance either. Programs I've coded do in fact run faster on the 1055T, but it all depends on what your doing.

[informal]

But PII X6 @ 4GHz is not as good as i7 quad core at same clock in majority of cases



I wonder why people compare AMD 6 core to intel 6 core ?

Only the last 4 are the actual real workloads,2 of them being games with poor MT optimizations and the other 2 showing almost identical performance(where MT optimization is present and done properly).Remember that Nehalem core does have IPC advantage in single threaded applications,it varies from none to 15-20% but it's present.Nobody is denying that. One may argue this advantage is not that big(I'm the one) and it all boils down to price/perf./watt ratio. Basically you get 2 relatively evenly matched systems that each has its strong points.It depends on what you are using your system for and how much you get for the $ you pay.

[gosh]

Even using linux, where the performance difference is minimal, there are noticeable differences in performance between the 1055T and the Core i7 920. Compilation time for software is still faster on the 920 (5-7%), and sequence alignment using indexing and a Wheeler-Burrows algorithm is still faster (about 8% or so, software is called Bowtie for those interested). The 1055T is no slouch for sure, but it isn't a complete 920 killer in every instance either. Programs I've coded do in fact run faster on the 1055T, but it all depends on what your doing.

Of course it isn't a 920 killer but in general you get more for same amount of money.
Visual Studio does perform better on X6, compiling more than one project and more cores is good to have

[Computurd]

dont wanna wade into the ol amd v intel fray at all, but i am seriously disappointed at the price point of this chip. i was expecting less. much less. i guess sandy bridge enthusiast class will be my first 6 (or 8) core.

[FlanK3r]

yes...NB, performance of Phenoms.

[zelos]

We can not compare amd and intel on this 6 cores categuory.
Even if amd has a 6 cores, intel is far away with the HT technology.
The amd 6cores can be compare to the i7 920 ( HT on equivalent to 6 physicals cores on perf ).
We have to wait the amd HT to compare the same cores.

[OutSider]

and still does not change the fact that its not cheapest 6core

[duploxxx]

But PII X6 @ 4GHz is not as good as i7 quad core at same clock in majority of cases



I wonder why people compare AMD 6 core to intel 6 core ?

why compare them if you clock them the same? do the test again @ stock... if you want to oc anyhow you can just order a T1055 and oc it.

We can not compare amd and intel on this 6 cores categuory.
Even if amd has a 6 cores, intel is far away with the HT technology.
The amd 6cores can be compare to the i7 920 ( HT on equivalent to 6 physicals cores on perf ).
We have to wait the amd HT to compare the same cores.

then you will have to wait for a very very long time, perhaps never
it doesn't matter if you make a design for ht or more cores, logic is much bigger on a core with HT then without, this is a development choice.
If current AMD would have the same IPC/ghz Intel would be in very bad shape....with or without HT. But we know k10 has lower IPC so it will be depending on BD, sandy is just again an enhancement, not a real leap forward, this was done with Nehalem.

[Scubar]

no...
this is the cheapest six core


http://www.newegg.com/Product/Produc...-851-_-Product

Yea but that one will be absolute pants in comparison to the Intel offering. Even the i7s out already trample on it.

[dartaz]

why compare them if you clock them the same? do the test again @ stock... if you want to oc anyhow you can just order a T1055 and oc it.

At stock i7 930/860 match the performance of T1090

If you overclock them to the same clock, the i7 is better in most cases

[pokipoki]

Don't put too much hope on benchmarks, some of them are rigged (benchmarketing) by the compilers (intel), processors which doesn't carry the GenuineIntel cpuids' will be compiled to run in non-optimized code, thus the lower benchmark scores & performance in some apps. It's part of the anti-trust case against Intel.

Be an informed user and buyer.

http://www.vanshardware.com/articles...el_SysMark.htm
http://www.osnews.com/story/22683/In...from_Compiler_
http://arstechnica.com/hardware/revi...o-review.ars/6
http://www.brightsideofnews.com/news...substance.aspx

In any case, it is relatively easy to test for CPUID vendor name string sensitivity. PCMark05, SysMark 2007 Preview, and Everest Ultimate are three benchmarks that have been affected by the vendor name string value that CPUID returns, but others exist.

How many benchmarks are run using the same cpuid for different brands of cpu?

[OhNoes!]

Don't put too much hope on benchmarks, some of them are rigged (benchmarketing) by the compilers (intel), processors which doesn't carry the GenuineIntel cpuids' will be compiled to run in non-optimized code, thus the lower benchmark scores & performance in some apps. It's part of the anti-trust case against Intel.

Be an informed user and buyer.

http://www.vanshardware.com/articles...el_SysMark.htm
http://www.osnews.com/story/22683/In...from_Compiler_
http://arstechnica.com/hardware/revi...o-review.ars/6
http://www.brightsideofnews.com/news...substance.aspx



How many benchmarks are run using the same cpuid for different brands of cpu?

Uh oh!, you're opening a can of worms. It's simple, really:

Run vendor agnostic benchmarks or even real world apps on linux and windows that have no such compiler code and the results are relatively same. If you're in doubt, check SPEC results for AMD vs Intel server processors, the results are replicated on the desktop.

[pokipoki]

Uh oh!, you're opening a can of worms. It's simple, really:

Run vendor agnostic benchmarks or even real world apps on linux and windows that have no such compiler code and the results are relatively same. If you're in doubt, check SPEC results for AMD vs Intel server processors, the results are replicated on the desktop.

It's for those who believe the benchmarks so much. No one want to admit they've been had, but to have it done to them again and again begs the question : How does it feel to line up the coffers of those who cheats you? Read the latest Intel earnings report?

And these are the smart and confident consumers, who advises others to be like them?

sheesh.

[OhNoes!]

It's for those who believe the benchmarks so much. No one want to admit they've been had, but to have it done to them again and again begs the question : How does it feel to line up the coffers of those who cheats you? Read the latest Intel earnings report?

And these are the smart and confident consumers, who advises others to be like them?

sheesh.

Hmmm? Again you have a point IF SPEC benchmarks were any different from all the review results we've been seeing. The issue you're raising has been beaten to death, do a search; besides, the salient point you're ignoring is that some of these things you call "cheats" are actually "optimizations" for optimum performance based on processor family capabilities. A crude example, you cannot 'tell' a P4 northwood processor to execute SSE 4.2 code when you don't know if it's capable of executing the code. So Intel uses flags to query each processor's id. and execute optimizations accordingly. So even some Intel processors get the same treatment as 'non-optimized-for" processors. These optimizations have to be tested and validated in Intel labs, etc. before they are implemented in code to prevent problems. Are you suggesting Intel labs should dedicate their time and money to optimize code for all rival processors in their compilers?And if Intel allows optimized code to run on rival processors untested, when a problem arises, who would provide support? Do you send your NVidia hardware software related questions to ATi for an answer? It's a complicated situation, over-blown, and misunderstood/misrepresented. Yet the answer is always there - test in a neutral environment and buy your hardware based on the results. Fair enough?

This is a non-issue.

[Wesker]

It's for those who believe the benchmarks so much. No one want to admit they've been had, but to have it done to them again and again begs the question : How does it feel to line up the coffers of those who cheats you? Read the latest Intel earnings report?

And these are the smart and confident consumers, who advises others to be like them?

sheesh.

Oh please, don't pretend that AMD is a charity. This is what seriously derails Intel/AMD threads, IMO. The fact that many AMD fans believe that AMD is some kind of church run foundation, saving the earth or something.

Intel offers a faster processor, period. BMW offers faster cars than Toyota, despite the huge price hike--people are still going to buy a BMW. Simple as that. You pay extra for a performance advantage.

Also, as OhNoes! stated, look at the SPEC benchmarks.

[DTU_XaVier]

Here in Denmark, you can get a 1090T for the same money as an i5-750... That tells some of the story... AMD is currently not trying to compete with Intel on the pure performance of their products, we all know that Intel is, at current, faster... But it's hard to argue with the performance to price ratio of their current lineups, even if they ARE just trying to scrape by while Intel holds onto the crown like a greedy tyrant (not implying that Intel is a monopolistic cashhoarder, but they do run a business)...

Best Regards

[Blacky]

$900 at newegg while 980x is at $999....