Another member of the Haswell club

[cytg111]

As the old rig failed all and about, i bit the bullet and got some new parts

- 4770(non k)
- Asrock extreme6 (multi core turbo, and i could decide to go K later on)
- 32G Kingston 1600MHz RAM
- Corsair GS800 psu
- Gigabyte windforce gtx780

....... Wonder if it will play Crysis ? Just got it up and running.. smooth experience.

 

[frozentundra123456]

As the old rig failed all and about, i bit the bullet and got some new parts

- 4770(non k)
- Asrock extreme6 (multi core turbo, and i could decide to go K later on)
- 32G Kingston 1600MHz RAM
- Corsair GS800 psu
- Gigabyte windforce gtx780

....... Wonder if it will play Crysis ? Just got it up and running.. smooth experience.

Nice!!

Can you overclock the non-k to the highest turbo for all cores with that MB? Not sure if you can do that still on Haswell.

 

[cytg111]

Nice!!

Can you overclock the non-k to the highest turbo for all cores with that MB? Not sure if you can do that still on Haswell.

http://www.anandtech.com/show/6989/...aswell-gigabyte-msi-asrock-and-asus-at-200/24

"In terms of performance the Extreme6/AC comes with MultiCore Turbo as standard, giving that extra boost over motherboards that do not have it."

.. Sure hope it does, its the reason I picked it over Asus when going non-K

 

[Kenmitch]

Nice rig

My Gigabyte board has a setting something like N OC? that says it'll do it also. I have a 4670k so no way to see.

Test it and see if it works I guess is the only way to know.

 

[AustinInDallas]

Why did you go non-k

 

[cytg111]

Nice rig

My Gigabyte board has a setting something like N OC? that says it'll do it also. I have a 4670k so no way to see.

Test it and see if it works I guess is the only way to know.

- Ok, Ill look into it

Why did you go non-k

Going non-k, and this setup, I get 3.4GHz and 3.9GHz turbo + multicore turbo (that could translate into 2-3 cores at 3.9 or close to it).. Compare that to the average overclocking results on the K version here .. hitting what? 4.2 or 4.3. Thats a 10% overclock that you have to invest in much better cooling for too AND miss out on VT-d and TSX .. do a search on XEN and VT-d, people are actually having their cake and eating it too (*nix and game-doze running side by side) and TSX is something i wanna mess with on the code level when I get the time. So those 10% extra Hz didnt pass my math test.

btw .. running prime95 now for 15-20 minutes and core temps are 90-97 .. Im hitting throttling temps on stock cooling ... DAAAAMN.

Stopped it and temps are dropping like gravity hit em .. back to 50 and sub in 10 seconds..

- actually got a "LOG" on one core .. meaning it got throttled.

Of course, am running open case right now, closed case will mean better airflow from the fans.

 

[ShintaiDK]

Remember with VT-D that you need mobo support, else it wont work.

 

[PPB]

You can still up the BCLK to ~105 and get around 4.1 OC, then proceed to laugh at anyone that paid more for a K variant just to find out they cant break past 4.3ghz

 

[cytg111]

Remember with VT-D that you need mobo support, else it wont work.

I did some quick searches before settling on the Asrock

http://wiki.xen.org/wiki/VTd_HowTo

"Intel Z87, H87, Q87 and B85 (information provided by ASRock support) "

Cause at first it seemed like you needed a Q level chipset.. Apparantly not so.

 

[cytg111]

You can still up the BCLK to ~105 and get around 4.1 OC, then proceed to laugh at anyone that paid more for a K variant just to find out they cant break past 4.3ghz

-Laffin' aint nice, but the math is the same .. the k is just not worth it. Not as much in price but in lagging features.

edit : note, this PSU is actually dead silent, if not running prime or gaming the fan is not spinning.. Might look into a megahelm kind of cooler for the haswell to go for an ultra silent rig ... i like it so far.

 

[Arkaign]

Hitting throttle temps on stock speeds with a stock cooler is unacceptable imho. Are they still using those horrendous push-pin ludicrously horrible stock coolers?

I know Intel has the undisputed performance crown in CPUs, but I feel like their dominance has lent them a certain arrogance in terms of their desktop decisions. Crap TIM, crap coolers, missing SKUs (where is i3-K @ $100ish for the value crowd?), super mediocre performance improvements (eg; I'd probably get better performance in almost everything with my existing 5Ghz 2700K vs. 4770), late releases (where's IB-E? At this point we should be looking for Haswell-E, why so long to release the workstation-level parts?) .. argh.

I even went through several 3770K chips / boards before deciding that I didn't want to de-lid, and the temps and low OC ceiling was unacceptable to me. Now 4xxx looks like a skip to me as well, though I can understand the lure if someone can truly make use of the new instructions. I don't code, and I almost never encode anything, so I truly don't see a value proposition with IB or Haswell.

 

[parablooper]

Hitting throttle temps on stock speeds with a stock cooler is unacceptable imho. Are they still using those horrendous push-pin ludicrously horrible stock coolers?

I know Intel has the undisputed performance crown in CPUs, but I feel like their dominance has lent them a certain arrogance in terms of their desktop decisions. Crap TIM, crap coolers, missing SKUs (where is i3-K @ $100ish for the value crowd?), super mediocre performance improvements (eg; I'd probably get better performance in almost everything with my existing 5Ghz 2700K vs. 4770), late releases (where's IB-E? At this point we should be looking for Haswell-E, why so long to release the workstation-level parts?) .. argh.

I even went through several 3770K chips / boards before deciding that I didn't want to de-lid, and the temps and low OC ceiling was unacceptable to me. Now 4xxx looks like a skip to me as well, though I can understand the lure if someone can truly make use of the new instructions. I don't code, and I almost never encode anything, so I truly don't see a value proposition with IB or Haswell.

Your opinion has been stated by many others on many occasions... why even bother?

It's widely agreed upon that Haswell is a sidegrade for desktops, especially enthusiasts. The whole "K series" is just a money scam now, and overclockable i3's are not gonna happen.

Ivy Bridge-E will release September 1st, and I'm looking forward to the 4820K. Haswell i3s release on the same day.

Our only hope is that when Intel gets around to working on Haswell-E, they'll be like, "Hey, this is a real shit architecture to work with when we're trying to get performance, especially performance/power." To prevent releasing an AMD-esque 220W TDP line, hopefully they'll realize the error of their ways, and start a series of true enthusiast processors, soldering and all.

But they won't. The money is in tablets. Touch-screens. 4.5W TDPs with no heatsink needed. Haswell was a great improvement for Ultrabooks and the like. That's the direction they're going to go.

 

[Sheep221]

Hitting throttle temps on stock speeds with a stock cooler is unacceptable imho. Are they still using those horrendous push-pin ludicrously horrible stock coolers?

I know Intel has the undisputed performance crown in CPUs, but I feel like their dominance has lent them a certain arrogance in terms of their desktop decisions. Crap TIM, crap coolers, missing SKUs (where is i3-K @ $100ish for the value crowd?), super mediocre performance improvements (eg; I'd probably get better performance in almost everything with my existing 5Ghz 2700K vs. 4770), late releases (where's IB-E? At this point we should be looking for Haswell-E, why so long to release the workstation-level parts?) .. argh.

I even went through several 3770K chips / boards before deciding that I didn't want to de-lid, and the temps and low OC ceiling was unacceptable to me. Now 4xxx looks like a skip to me as well, though I can understand the lure if someone can truly make use of the new instructions. I don't code, and I almost never encode anything, so I truly don't see a value proposition with IB or Haswell.

yeah it's crap, well intel is only good because AMD is worse that's all

 

[cytg111]

Hitting throttle temps on stock speeds with a stock cooler is unacceptable imho. Are they still using those horrendous push-pin ludicrously horrible stock coolers? ....
(where's IB-E? At this point we should be looking for Haswell-E, why so long to release the workstation-level parts?)

- I agree, I didnt expect to get throttled at stock I must admit, and if it wasnt for my old rig craaping out I was well into giving the old q9450 another year of duty.... looking forward to haswell-e, given the current circumstances.
Oh well, now this is my new 5 year rig (on the memory part at least).

 

[Kenmitch]

- I agree, I didnt expect to get throttled at stock I must admit, and if it wasnt for my old rig craaping out I was well into giving the old q9450 another year of duty.... looking forward to haswell-e, given the current circumstances.
Oh well, now this is my new 5 year rig (on the memory part at least).

The stock cooler sucks and isnt worthy of your rig.

With my 4670k running stock settings vcore on auto my cbjp asks for .1v more under heavy stress testing than it requires to be stable at 4.2ghz.

Maybe try taking controll of your vcore.

 

[Essence_of_War]

The stock cooler sucks and isnt worthy of your rig.

Truth.

Blow $30 on a CM EVO, or equivalent. That i7 will thank you.

 

[inf64]

I don't get how can they sell chips that throttle at stock with stock cooler? Why don't they just bundle something better in terms of cooling?

 

[blackened23]

I don't get how can they sell chips that throttle at stock with stock cooler? Why don't they just bundle something better in terms of cooling?

The entire issue is that the Haswell's iVR over shoots the voltage when using stress test programs. This will cause a 1.1V stock voltage to shoot up to 1.2xV which obviously causes problems.

The Haswell chip will not throttle under normal usage with a stock cooler. Stress testing is a different story because of the aforementioned issue - the integrated voltage regular does not play correctly with prime95. The voltage auto corrects to an excessively high value which skews the intent of the test. On some boards you can use manual voltage to prevent the over-corrections (doesn't work on ALL boards though), but really. Why worry about prime95 at stock settings. You wont' have issues in real world usage, and it should not throttle.

My IVB reaches near 90C in prime 95, and outside of prime95 it has never exceeded 65C. Even during excessively high CPU workloads. I don't think the OP will have any issues with throttling when he's actually using his PC normally. (ie NOT prime95/IBT)

 

[PPB]

The question is, why is shooting +0.1v on AVX loads? Would it be unstable if it wasnt for that extra pump in voltage on those kind of loads?



Its a pretty big deal a stock cooler can't handle stress testing loads without throttling, because even if normal use wont make you reach those kind of temps, I know a program or two that, when used at full load, they make your cpu reach temps that are dangerously close to P95's temps. Yep 3d rendering can stress your system that much.

 

[cytg111]

The entire issue is that the Haswell's iVR over shoots the voltage when using stress test programs. This will cause a 1.1V stock voltage to shoot up to 1.2xV which obviously causes problems.

The Haswell chip will not throttle under normal usage with a stock cooler. Stress testing is a different story because of the aforementioned issue - the integrated voltage regular does not play correctly with prime95. The voltage auto corrects to an excessively high value which skews the intent of the test. On some boards you can use manual voltage to prevent the over-corrections (doesn't work on ALL boards though), but really. Why worry about prime95 at stock settings. You wont' have issues in real world usage, and it should not throttle.

My IVB reaches near 90C in prime 95, and outside of prime95 it has never exceeded 65C. Even during excessively high CPU workloads. I don't think the OP will have any issues with throttling when he's actually using his PC normally. (ie NOT prime95/IBT)

- Just had a session of "Metro: Last night". I was popping out regularly to check temps and they were in the 60-70 celcius range.. nowhere near prime95.. Also checking taskmanager it is clear that the game only tasks the real cores, no real hyperthreading going on. - I guess that is one "real world scenario".

 

[Idontcare]

The entire issue is that the Haswell's iVR over shoots the voltage when using stress test programs. This will cause a 1.1V stock voltage to shoot up to 1.2xV which obviously causes problems.

The Haswell chip will not throttle under normal usage with a stock cooler. Stress testing is a different story because of the aforementioned issue - the integrated voltage regular does not play correctly with prime95. The voltage auto corrects to an excessively high value which skews the intent of the test. On some boards you can use manual voltage to prevent the over-corrections (doesn't work on ALL boards though), but really. Why worry about prime95 at stock settings. You wont' have issues in real world usage, and it should not throttle.

My IVB reaches near 90C in prime 95, and outside of prime95 it has never exceeded 65C. Even during excessively high CPU workloads. I don't think the OP will have any issues with throttling when he's actually using his PC normally. (ie NOT prime95/IBT)

Prime95 is not a power-virus program though. Prime95 is an actual program that people use to find Mersenne primes.

Intel should not have released haswell if the iVR algo's are this poorly tuned, or if haswell itself has a critical speedpath in its AVX circuits that necessitates such an elevated voltage in order to reliably operate at the target clockspeeds.

To me this reeks of an immature product, which either means a new stepping is in the offing that addresses the gating speedpath in the AVX circuits (and folks are advised to hold out for it before upgrading to Haswell) or it means Intel really phoned it in on Haswell's iVR (ala Silverthorne style) which would justifiably lead to all sorts of questions regarding the legitimacy of their TDP spec (no different than the scrutiny leveled against AMD and the FX-8350's TDP).

 

[PPB]

If it's the first case, they have 2 years for a new stepping a la i7 9xx D0.

 

[ShintaiDK]

I don't get how can they sell chips that throttle at stock with stock cooler? Why don't they just bundle something better in terms of cooling?

It doesnt. And we already covered this with IB as well in IDCs old thread. When Mobo makers autooverclocks turbobins, it happens. The turbobins for stock is 4/3/2/2, not 4/4/4/4.

Asrock extreme6 (multi core turbo, and i could decide to go K later on)

Its amazing how fast people forget.

 

[inf64]

Remember this is not the 1st time we hear about this issue . It happened on other boards that do not feature MC Turbo or other similar gimmicks to artificially boost out of the box performance.

 

[ShintaiDK]

Remember this is not the 1st time we hear about this issue . It happened on other boards that do not feature MC Turbo or other similar gimmicks to artificially boost out of the box performance.

We did? Can you link it? Because every single time I seen it its been due to mobo autoOC.