Solved Haswell RAM issue

[ICIT2LOL]

John, nothing to fret over, the problem is with memtest, not Haswell. Intel has changed the innards of Ivy and Haswell so that most of the normal benchmarks don't have anything to read from. From what I understand, Haswell handles RAM much better in different ways but isn't reflected in many tests. Memtest only check for errors, it can't tell us how efficient RAM is.

Ok Gary I won't try the memtest I am interested in finding that new memtest though.

 

[Layback Bear]

Gary a couple of thoughts.
1. Check your temp across your cpu cores and see how close to each other they are.
2. Don't use XMP and see what happens.
Note: On my Asus motherboard I have a choice.

 

[Britton30]

@ John: The memtest 86+ version 4.2 is the newest. There is a memtest 86 (no +) but is not better.

Jack, not using the XMP results in the same readings and the RAM defaults to 1600, 11-11-11. My core temps are within 7C of each other.

 

[kbrady1979]

Was your winsat mem score any different on the Haswell rig?

 

[Britton30]

Yes Kelly, look here: http://www.sevenforums.com/overclocking-case-mods/285427-show-us-your-rig-5-a-7.html#post2483109
The Sandy rarely gets over 29,000.

 

[kbrady1979]

OK, thanks. I knew you had it posted somewhere, but couldn't remember where. I'm just going to assume that even though the benchmarks don't reflect it, Haswell utilizes RAM differently and possibly more efficiently.

 

[Britton30]

Yes, that's my take on it too.

 

[Layback Bear]

Are you still using the stock Intel cooler?
I don't know about your Haswell chip but for other Intel chips 7C is a lot of difference. I would watch them closely. If the chip get hot the memory controller get hot.

 

[kbrady1979]

I thought 7* difference under load was pretty good actually. Ivy Bridge has some variation but Sandy Bridge had quite a bit, sometimes well over 10*C

 

[Britton30]

I have a Cooler Master 212EVO on it now, dropped idle temps from 38C or so to the mid 20's. I first used just one fan then added the second one and made zero difference in temp, idle and load.

My Sandy's temp spread is 16C under IBT. I think it's because I let it get to 100C once. Otherwise it still runs well.

Oh, I did run Prim95 14.5 hours on Haswell at 4.2GHz with no errors and max temp was 69C, I'm happy with that.

 

[kbrady1979]

I have a Cooler Master 212EVO on it now, dropped idle temps from 38C or so to the mid 20's. I first used just one fan then added the second one and made zero difference in temp, idle and load.

My Sandy's temp spread is 16C under IBT. I think it's because I let it get to 100C once. Otherwise it still runs well.

Oh, I did run Prim95 14.5 hours on Haswell at 4.2GHz with no errors and max temp was 69C, I'm happy with that.

That's pretty damn good Gary!

 

[paulpicks21]

Nice temp on the prime Gary, I think they run quite similar to Ivy as far as temps go, you get to around 4.2-4.4GHz fine and then the temps go mental.

 

[kbrady1979]

I never understood what people were talking about with IB having crazy temps. It was more sensitive to voltage, but I've never seen any crazy temps out of mine. I think there were some bad CPU's out there, and maybe that is where all the talk came from...idk

 

[paulpicks21]

I never understood what people were talking about with IB having crazy temps. It was more sensitive to voltage, but I've never seen any crazy temps out of mine. I think there were some bad CPU's out there, and maybe that is where all the talk came from...idk

Sorry Kelly, maybe 'Mental' was a little over the top, I think I was just too used to the temps of Sandy and then when I got Ivy the temps were considerably higher at the same speeds, but with Ivy it is perfectly fine as they do run hotter and obviously the TJ Max is 105. I did manage to hit that at 5GHz lol.

Before I was used to highs of mid 60s to 70s and with Ivy I have had to get used to temps much higher than that when running prime and IBT.

I worded it wrong.

 

[Layback Bear]

Very good temps Gary.
Although a 960 Bloomfield do run hot I have never had over 2 C difference in core temps. With any cpu you will find a voltage point where the temps make a major jump. That is why I mentioned watching the temps.
Some where a long time ago Intel stated that idle temps could be off quiet a bit because their sensor were more calibrated for high temp.
They were not concerned how low temps got at idle but how hot they got under load. That makes sense. Full load at 4.2 and 69C is great temps.
My low temps will be about 10 to 12c lower in bios compared to program temps. Whether the 10 to 12 degrees difference carry on out to the hight load temps; I don't know and have no way of checking.
What is your vcore at 4.2?

 

[kbrady1979]

I never understood what people were talking about with IB having crazy temps. It was more sensitive to voltage, but I've never seen any crazy temps out of mine. I think there were some bad CPU's out there, and maybe that is where all the talk came from...idk

Sorry Kelly, maybe 'Mental' was a little over the top, I think I was just too used to the temps of Sandy and then when I got Ivy the temps were considerably higher at the same speeds, but with Ivy it is perfectly fine as they do run hotter and obviously the TJ Max is 105. I did manage to hit that at 5GHz lol.

Before I was used to highs of mid 60s to 70s and with Ivy I have had to get used to temps much higher than that when running prime and IBT.

I worded it wrong.

I knew what you meant, but I've heard others talk about the extreme temps of Ivy Bridge. I think what messed people up was that Sandy Bridge you could throw the voltage to it and the temps were likely to be ok and Ivy Bridge was more sensitive to higher voltages causing temps to sometimes get out of hand.

 

[paulpicks21]

I never understood what people were talking about with IB having crazy temps. It was more sensitive to voltage, but I've never seen any crazy temps out of mine. I think there were some bad CPU's out there, and maybe that is where all the talk came from...idk

Sorry Kelly, maybe 'Mental' was a little over the top, I think I was just too used to the temps of Sandy and then when I got Ivy the temps were considerably higher at the same speeds, but with Ivy it is perfectly fine as they do run hotter and obviously the TJ Max is 105. I did manage to hit that at 5GHz lol.

Before I was used to highs of mid 60s to 70s and with Ivy I have had to get used to temps much higher than that when running prime and IBT.

I worded it wrong.

I knew what you meant, but I've heard others talk about the extreme temps of Ivy Bridge. I think what messed people up was that Sandy Bridge you could throw the voltage to it and the temps were likely to be ok and Ivy Bridge was more sensitive to higher voltages causing temps to sometimes get out of hand.

Yeah with Sandy it was a gradual voltage/temp curve that could be predicted, with Ivy, once you went above a certain speed the voltage increases and temp increases were quite substantial and surprised a lot of people me included.

 

[paulpicks21]

Sorry Gary, this is off topic, I think me and Kelly have done this twice tonight now :roflmao:

I will refrain from further comments.

 

[Layback Bear]

A good little artical on Haswell. I doesn't give a step by step but a very good idea how it all works together.

3 Step Guide to Overclock Your i7 / i5 Haswell Platform | Overclockers

 

[Britton30]

Sorry Gary, this is off topic, I think me and Kelly have done this twice tonight now :roflmao:

I will refrain from further comments.

That's OK with me, we all learn that way.

I think the volts reported in CPUz at 4.2GHz is 1.207. I don't know if that's Vcore or not, there are so many terms. As a comparison the Sandy at 4.4GHz runs 1.344-1.368V.

On the Sandy 4.4GHz again, here is Core Temp read running IBT standard, note the !5C spread in core temp.