Intel 22nm 'Ivy Bridge' Core i7 3770K Reviews

[zalbard]

and then how did they get 4.8ghz on air with IB while the rest of the world seemed stuck at 4.5-4.6

Most got to 4.7-4.9 before becoming temperature constrained.

[SKYMTL]

i checked it out and i do not like the OC results for power consumption.
4.8ghz on SB should not take 1.48v
and then how did they get 4.8ghz on air with IB while the rest of the world seemed stuck at 4.5-4.6

I didn't see mention of their IVB voltage either.

[Eldonko]

Was thinking the same thing. There is a whole section..

Read ours. We dinged it no less than 10 times in our article, and with good reason.

[hokiealumnus]

Read ours. We dinged it no less than 10 times in our article, and with good reason.

Another +1; ours talked about it as well.

[Manicdan]

I didn't see mention of their IVB voltage either.

its at the end. 1.28v

EDIT, 1.27v sorry

[Charles Wirth]

Any iGPU overclocking articles?

Mike and I both did iGPU overclocking. UD3 and Gene V

[R101]

Mike and I both did iGPU overclocking. UD3 and Gene V

Thanks, I will check it out. Quick sync is more important to me than 5-10% cpu speedup.

[Boogerlad]

I don't understand the purpose of quicksync. It has worse quality and speed compared to x264 --ultrafast preset.

[Trox]

Did you do any iGPU + memory overclock bencharking?

Plase share the link,

Thanks,

bruno

[tbone8ty]

i think semiaccurate did an article on igpu overclocking, conclusion it didnt gain much

id be more interested if somebody donated their cpu to the cause and took the heat spreader off and then OC'd it see if that helps with heat disapation

[kl0012]

I don't understand the purpose of quicksync. It has worse quality and speed compared to x264 --ultrafast preset.

According to Anand, QS is 2x faster then Handbrake with acceptable image quality.
http://www.anandtech.com/show/5771/t...770k-review/21

[Boogerlad]

I don't see any screenshots of the x264 encode. And where is the original pre encoded screenshot?

[bamtan2]

I don't understand the purpose of quicksync. It has worse quality and speed compared to x264 --ultrafast preset.

I dont understand the purpose of this post with no benchmarks and proof

[BlindFreddie]

id be more interested if somebody donated their cpu to the cause and took the heat spreader off and then OC'd it see if that helps with heat dissipation

+1 on the IHS exercise

I think:

1. There must be a bigger-than-usual thermal resistance bottleneck between the IB die and the IHS. IB apparently draws less power than SB at the same clock, yet its on-board thermal sensor says it runs hotter with the same cooling setup. Assuming that the thermal sensors have the same accuracy, the laws of physics say that for this to happen, our external heatsink must run cooler when on IB. Has anyone checked? (Putting an ordinary thermometer in the airstream downwind of the HSF would tell us).
2. [[I have an idea that]] Someone posted in another thread that the IB chip is not soldered to the IHS and uses some yucky old TIM instead ([Edit]: http://www.xtremesystems.org/forums/...preader-Report[/Edit]), in which case it may be comparatively easy to remove the top of the IHS. Hopefully you wouldn't need to electrically insulate the top of the chip from the external HS that you'd then install.
   Maybe try removing the IHS from a dead chip; next level would be to do it with the cheapest IB chip available.
   If the problem is due to the surface of the silicon being microscopically rough instead of flat as with conventional 2-D transistors, removing the IHS may not help much though.
   Sorry, but I don't have an IB chip dead or alive, or a sponsor, or a Newegg or a Microcenter or anyone else who will sell me an IB at a cheap price.

@hokiealumnus: Thanks for your review. Re the temperature of IB vs SB, what no-one else has suggested is that if you want to state a % difference, you should be talking about the rise above ambient. That eliminates the problems with quoting above absolute zero, and also cancels out the differences between the Fahrenheit and Celsius scales.

And those of us who are looking at these chips not for gaming, benching or personal status, but to contribute to grid computing projects such as worldcommunitygrid haven't had much of a say here yet. We need high CPU processing output per watt and long CPU life - up to 5 years - at the highest speed that is compatible with those constraints. It's disappointing that IB does not seem to enable that speed to be pushed much higher if at all, but the saving in power consumption is significant.

I'll wait a month or 2 to see whether bugs appear in the new architecture or whether any improved steppings come along, then upgrade to IB. I have a very sick 2600K that is hanging onto life by a whisker at 4.2GHz and its max Vcore without further degradation of 1.23V and memory at its max stable clock of 1066MHz , plus a couple of 45nm C2Qs that are great machines but they've gotta go, in the face of imminent huge increases in electricity costs here.

[bamtan2]

blindfreddie, it is hotter with less voltage because the chip is smaller. the IHS is probably really well made and fastened on. it is just a curiosity that removing it and directly connecting to something else might help.

[R101]

I don't understand the purpose of quicksync. It has worse quality and speed compared to x264 --ultrafast preset.

I use 2600K (about 30 of them) right now.. You can check the quality of the encoding in the Anandtech original SB article.

The speedup that IB brings to the table (at least in my usage scenario) is just amazing. And Haswell cannot come soon enough .

[Leeghoofd]

i think semiaccurate did an article on igpu overclocking, conclusion it didnt gain much

That's why they are called Semiaccurate :p I got good gains with IGP OCing and faster Ram (2600 vs 1600Mhz) in some games...

[Olivon]

Intel Core i7-3770K et i5-3570K : Ivy Bridge 22nm - Hardware.fr



Very interesting, as usual

[kaktus1907]

old bartons were 101mm2 and 75W doesnt have the problems like ivy.. i could make it stay cool under 2500MHz like 65-70C.. it should be near drawing 150W or so and even smaller diesize than ivy relatively..

[Cold Fussion]

old bartons were 101mm2 and 75W doesnt have the problems like ivy.. i could make it stay cool under 2500MHz like 65-70C.. it should be near drawing 150W or so and even smaller diesize than ivy relatively..

With much lower currents though.

[kaktus1907]

With much lower currents though.

that's my point.. i dont buy what everyone says oh it's such a smaller die size that hsf cant cool it blah blah.. we had/have much smaller and similar power draw consumption numbers various chips that this wasnt the case until ivb but now everyone says that..

[Olivon]

- Sandy Bridge : 0.87 watts per mm²
- Ivy Bridge : 1.11 watts per mm²

The decline in consumption is not large enough compared to the size reduction of core, Ivy Bridge is more difficult to maintain in temperature.

http://www.hardware.fr/articles/863-10/temperature.html

[kl0012]

old bartons were 101mm2 and 75W doesnt have the problems like ivy.. i could make it stay cool under 2500MHz like 65-70C.. it should be near drawing 150W or so and even smaller diesize than ivy relatively..

This comparison doesn't make sense. Bartons didn't have precise integrated temperature sensors, so you don't really know what temperatures happens inside the chip.

[hokiealumnus]

@hokiealumnus: Thanks for your review. Re the temperature of IB vs SB, what no-one else has suggested is that if you want to state a % difference, you should be talking about the rise above ambient. That eliminates the problems with quoting above absolute zero, and also cancels out the differences between the Fahrenheit and Celsius scales.

Thank you for that; and you're absolutely correct; I didn't mention that but absolutely should have. Ambient was very close to the same between the two within two degrees Celsius. I'm kicking myself for not measuring when I took it, but that room stays very close to 23°C, ticking a degree or two one way or the other (except in the middle of summer, when it can heat up more).

[PatRaceTin]

http://www.hardware.fr/articles/863-10/temperature.html

Interesting information