7.1 ghz??? how come the ph2 955 can reach that high and not the i7?

[rooseveltdon]

I just read that a team of overclockers were able to push the phenom 2 955 all the way to 7.1 ghz what i dont understand is why i don't hear the i7 doing better,it's built on better technology i would expect it to be able to be pushed that high but i don't hear anything about the i7 going past 5ghz even on extreme cooling solutions. Why is that? please no flame wars between fan boys i am just curious.

 

[the last resort]

AMD has managed to eliminate the "cold bug" meaning that you can run the processor very cold, i.e. liquid helium at just over a degree Kelvin above Absolute Zero. Intel's Core i7 IS a better technology for performance; just because the AMD can overclock farther than the Intel i7 does not meant that the i7 is bad. In fact, IMO the Intel could probably beat that 7.1 GHz without being clocked nearly that high.

Its just that AMD has managed to design and build a processor that overclocks VERY well, and does provide a good amount of performance.

 

[mlcloud]

The i7 965 has been overclocked to 5.5ghz before, although I'm not sure if it had been confirmed.

And remember that the clock-to-clock performance ratio is different between the PII and i7 builds (IE, a Pentium 4 running at 3.0ghz != single core of a Core 2 Duo running at 3ghz), and for all intents and purposes, the i7 is better at usable, stable clock ranges. Although AMD does deserve some bragging rights for a 7.1ghz quad core...

 

[Wisecracker]

WooHoo! Let's play synthetics!

I don't really care what guys do on suicide runs. For the most part I think it encourages the bone-heads with small lives to argue 'mine's bigger than yours'.

The other 98.8% of the world could not care less. All they want is their email and p0rn - or maybe to email their p0rn.

Instead of over AMD/Intel people should bitch about all the wasted bandwidth from the 60%+ of useless email ...

 

[cybot_x1024]

It probably just overclocks better and thats just about it.

 

[rabidbunny]

The i7 965 has been overclocked to 5.5ghz before, although I'm not sure if it had been confirmed.

And remember that the clock-to-clock performance ratio is different between the PII and i7 builds (IE, a Pentium 4 running at 3.0ghz != single core of a Core 2 Duo running at 3ghz), and for all intents and purposes, the i7 is better at usable, stable clock ranges. Although AMD does deserve some bragging rights for a 7.1ghz quad core...

Um, isn't that the pentium dual core where a P4 @ 3.0ghz = one core of a pentium d? the C2D is WAY faster core for core than a P4.

 

[silversurfernhs]

"!=" is a programming code for "not" for anyone that didn't understand his equation.

 

[rabidbunny]

"!=" is a programming code for "not" for anyone that didn't understand his equation.

Oh, sorry! I didn't see the !. stupid of me.

 

[hdwenthus]

well i believe AMD has geared their processors to that end, while intel has not

 

[Zenthar]

"!=" is a programming code for "not" for anyone that didn't understand his equation.

In some languages it can also be "<>"

BTW, one thing I find interesting about this news is that it tells the AMD CPU can operate in rather extreme conditions. For most people it doesn't matter, but for people using handhelds in arctic and/or in space, it can matter.

 

[yomamafor1]

That's because, like most people have already said, Core i7 has a much more complicated architecture than Phenom II. Not only that, Intel chose a different approach when designing the transistors, which results in lower heat output at stock voltage.

As for clockings, Pentium 4 has been seen operating well in the 8Ghz range.

 

[JAYDEEJOHN]

I think its because Intels arch is wider. If you go wider, it usually means your IPC will be better, but it also usually shortens the pipes, and limits overall ocing to a degree.

 

[jimmysmitty]

I just read that a team of overclockers were able to push the phenom 2 955 all the way to 7.1 ghz what i dont understand is why i don't hear the i7 doing better,it's built on better technology i would expect it to be able to be pushed that high but i don't hear anything about the i7 going past 5ghz even on extreme cooling solutions. Why is that? please no flame wars between fan boys i am just curious.

Because when you use extreme cooling methods you can OC a lot of things to a quite high level. There is a CPU-Z record for a Pentium 4 running at 8GHz.

But the thing is that most Phenom IIs, even the 955, only OC to 3.8GHz on air and after that require either a ultra large heatsink or water cooling.

Core i7 was the same until the recent D0 stepping where a Core i7 920 is able to reach 4.3GHz on air.

Intels Core i5 32nm quads can reach 3.6GHz on air within the same thermal envelope meaning that possibly OCing them will be better and easier without having to resort to exotic cooling (peltier, water or compressed air).

Besides, we all know that these results are useless on normal Desktops. They are just for people who like to OC the bejebus out of things. One guy was able to take a HD2900XT and OC its core to 1.6GHz+ from the stock 700MHz. But he did that just to get the 3DMark06 record.

 

[someguy7]

How did this become a clock for clock thing?

All the op asked was how come the pII hit a higher max overclock than the i7. Not about the performance or anything. Just WHY one cpu can go clock higher with extreme cooling than the other. So much for non flame war.

 

[MU_Engineer]

That's because, like most people have already said, Core i7 has a much more complicated architecture than Phenom II.

Not really, they are much more the same than they are different. Both chips have four cores on a die, three levels of cache, an IMC, and a split core/L3+IMC frequency plane. The biggest difference between the two is probably Intel's use of SMT.

Not only that, Intel chose a different approach when designing the transistors, which results in lower heat output at stock voltage.

That's much more likely to be part of the reason than any "extra complexity" of the i7.. IIRC, Intel also added two more transistors to their cache cells for lower power consumption but at a loss of performance. Also, Intel's and AMD's manufacturing processes are quite different and it could just be that AMD's leads to better overclocking this time around. Intel's 65 nm was pretty good for overclocking while AMD's wasn't, and the tables could have very well turned.

As for clockings, Pentium 4 has been seen operating well in the 8Ghz range.

That was most likely because the P4 Cedar Mill that hit 8 GHz had over 30 pipeline stages, compared to the Phenom II's supposed 12 stages. It's a lot easier to get a chip with a mile-long pipeline to clock higher than one with a shorter pipeline. Look at what IBM is doing with the current POWER, selling it at 4.6 or 4.7 GHz. It's got a pipeline over 20 stages too, IIRC.

 

[someguy7]

And on another note. The highest possible GHZ either cpu can reach is pointless. Its just for bragging rights. Where it does come into play for 99 percent of the overclockers are how far it will go on water/air.

As for why the amd can get a higher extreme overclock...I'm also ponder how the AMD can hit higher then the Intel. And not in a who is better type of thing. As far as I know both cpus are made with similar goals. (4 cores, IPC. Thermals-clockspeed) so I do wonder how one gets further under extreme temps then the other. Pack in the p4 days we all knew Intel had made the p4 for high clockspeeds but that it just didnt work for them because of the thermals.

 

[jimmysmitty]

And on another note. The highest possible GHZ either cpu can reach is pointless. Its just for bragging rights. Where it does come into play for 99 percent of the overclockers are how far it will go on water/air.

As for why the amd can get a higher extreme overclock...I'm also ponder how the AMD can hit higher then the Intel. And not in a who is better type of thing. As far as I know both cpus are made with similar goals. (4 cores, IPC. Thermals-clockspeed) so I do wonder how one gets further under extreme temps then the other. Pack in the p4 days we all knew Intel had made the p4 for high clockspeeds but that it just didnt work for them because of the thermals.

Expanding on this, I would say whats even more interesting is how the original Phenom couldnt't OC near that high on LN2 cooling but the Phenom II can.

I mean they are basically the same core design, but the Phenom II has a few IPC improvements and bug fixes. Other than that it is a die shrink and not a entirely new chip.

Could it be possible that they don't have a cold bug or have yet to find one while Core i7 does have a cold bug?

 

[JAYDEEJOHN]

Besides having the wider arch, until AMD uses HKMG, we cant eliminate it as a possible oc limiter, as the fine tuning using HKMG could limit it in a particular thermal window, or, SOI is slightly better than bulk at these processes.
I think its a combination. HKMG lowers v, but also allows for higher thermal creation outside its design, thus limiting oc. SOI vs bulk, well unless its the 45nm process, bulk has always held its own, tho, going to a native quad may have limited Intel also.
We really need 1 more gen to actually see how this all pans out.

 

[jimmysmitty]

Besides having the wider arch, until AMD uses HKMG, we cant eliminate it as a possible oc limiter, as the fine tuning using HKMG could limit it in a particular thermal window, or, SOI is slightly better than bulk at these processes.
I think its a combination. HKMG lowers v, but also allows for higher thermal creation outside its design, thus limiting oc. SOI vs bulk, well unless its the 45nm process, bulk has always held its own, tho, going to a native quad may have limited Intel also.
We really need 1 more gen to actually see how this all pans out.

Until we see what Core i5 can do its hard to say. I mean Turbo Mode itself clocks it to 3.6GHz within the same TDP. That may mean higher clocks at less voltage on air.

 

[Reynod]

I think the i7 may not overclock so well as it's prefetch and cache are very tightly tuned to increase IPC and the transistors in the cache have a higher idle current.

The Phenom II cpu's have much looser cache timings and so the processor overclocks very well.

Still, you and I are not going to play a decent game for very long running on LN2 ... I don't like the prospect of the burns topping up the canister every 5 minutes and it is going to get very expensive on the nitrogen ... as it is an open loop cooling system.

 

[JAYDEEJOHN]

They need to create a super compressor, ala Freon, heheh, close that loop

 

[JAYDEEJOHN]

Theres less tie downs regarding i5, so less to have to control during oc. i5 may hit a tad lighter, but its going to be a killer chip.

 

[Reynod]

Doesn't make any economical sense to release it as Intel would still have tons of inventory of the core2 line.

i5 would kill it's own line and confuse the market.

Thats why there is no i7 product below the 920 - they would be stockpiling all of the silicon that didn't meet that frequency and it will be fused off (reduce the mem contoller channles and cache) to make i5's until that mask is ready ot go ... if it has a new mask.

 

[Helloworld_98]

I thought that i5 was going to be base locked aswell as multiplier locked for the lower end versions.

Maybe I'm wrong but I could've sworn I read that somewhere.

 

[apache_lives]

That P2 overclock - depends on if they had all 4 cores running, all cache enabled, all 4 cores @ that clock speed, and dependent on stability etc

An i7 does have more cache, is a "younger" design (1st gen IMC/CSI design, second gen 45nm?) and has a generally higher IPC compared to the P2, so the P2 may need 1+ ghz to beat an i7 at a maximum overclock and does have a newer stepping that was just released wether they have compared that rev. or not?

Any perhaps the theory is right with i7 being more of a server cpu then desktop hence performance/efficency/ipc over scaling (clock speed) - afterall in dual socket benchmarks those babys own AMD quad socket solutions.