Intel Haswell to be easier to overclock
Some leaked PDF files that have been shown on IDF make it clear that Haswell will be easier to overclock. Intel said improvements have been made to the way you can overclock Haswell, making the process similar to tweaking Sandy Bridge-E CPUs. Interesting to see BTW is that the core multiplier ratios will go up to 80 with Haswell, nice.
Haswell should be fairly similar to the say the Core i7 3820 (Sandy Bridge-E) as it will allow both multiplier based overclocking and base clock overclocking when multipliers are locked in. You will be able to tweak your baseclock by 5-7% at three different presents – 100MHz, 125MHz and 167MHz. The reason why Intel quotes only 5-7% variance at those presets is because too much BCLK changing can cause instability in the PCI-Express and DMI-PLL. It is good to know that you will be able to overclock the chips, not just the K versions. There will be three base frequencies to which a multiplier can be applied: 100 MHz, 125 MHz, and 166 MHz. Several uncore frequency ratios will adjust themselves to compensate, thus leaving their stability unaffected.
Haswell will get integrated voltage regulation too brininging each CPU an integrated VRM controller. Unlocked “-K” chips remain to have an as they can reach 8 GHz record (subzero cooling) as their base clock multiplier for the CPU cores are higher than on the others, of up to 80.0x for 100 MHz, up to 64.0x for 125 MHz, and up to 48.0x for 166 MHz.
Memory overclocking is going to be popular with Haswell. Intel is to offer support for 200 MHz steps up to 2.6GHz and 266MHz steps up to 2.66GHz on memory frequency. The maxed out memory data rate supported will be a nice 2.93GHz !
Intel Haswell is around the corner GT3e spotted - 04/12/2013 10:09 AM
Over at the Asain edition of Intel's IDF a photo was taken of a Haswell processor with the heatspreader removed. This is thwe GT3e version of the CPU, a quad-core model with on-board RAM. Thus far In...
Intel Haswell Xeons get up to 15 cores - 04/03/2013 08:41 AM
CPU World has shown details about Intel's upcoming Haswell-based Xeon processors, you can read it over here. Xeon E7 v2 will arrive in Q4 2013, while Xeon E5-2600 v2 will launch a quarter earlier. Xe...
Noctua provides free mounting kit upgrades for Intel Haswell - 03/19/2013 04:02 PM
Noctua is being classy again as they announced that owners of its older heatsinks that don't include a SecuFirm2 mounting system are able to order a free NM-i115x mounting kit upgrade to make their c...
Intel Haswell to arrive in low volume until USB 3.0 issue is fixed - 03/15/2013 08:40 AM
Intel will slowdown the ramp of its new Haswell platform until it has fixed a USB 3.0 issue in the Haswell platform that causes some relatively minor issues when a system wakes from S3 sleep mode with...
Intel halts their own branded motherboards - 01/23/2013 10:04 AM
Intel indicated that ovre the next three years they will stop manufacturing their own branded motherboards. Well, they actually purchase them from big ODM's themselves but yeah that production will ...
Moderator
Posts: 10696
Joined: 2008-11-01
I will wait for officiall benchmarks first to see how good they will overclock assuming they will be using fluxless solder this time and not that cheap thermal compound like ivy,to be honest i personally think they might have intentionally used that cheap thermal paste on ivy to limit it's oc potential so more consumers coming let's say from Sandy bridge will be more tempted to upgrade to the new platform and not seeing it as a side grade, just imagine an Ivy bridge running at 5ghz on air performing similarly with a Haswell at let's say 4.6ghz (just an example) that for me is more of a side grade then an upgrade,and limiting the mainstream cpu's to 4 cores with HT,i know that they have practically no competition from Amd apart from the integrated GPU
but i would like to see at least 6 core parts with HT,before upgrading again....
Just my thoughts.............
Nice conspiracy theory. The reason they used paste was because they didn't have a feasible way to use fluxless solder to bind the new smaller 22nm die with the heat spreader. It was a technical limitation at the time, and had nothing to do with cost savings.
Moderator
Posts: 6559
Joined: 2009-10-12
I will wait for officiall benchmarks first to see how good they will overclock assuming they will be using fluxless solder this time and not that cheap thermal compound like ivy,to be honest i personally think they might have intentionally used that cheap thermal paste on ivy to limit it's oc potential so more consumers coming let's say from Sandy bridge will be more tempted to upgrade to the new platform and not seeing it as a side grade, just imagine an Ivy bridge running at 5ghz on air performing similarly with a Haswell at let's say 4.6ghz (just an example) that for me is more of a side grade then an upgrade,and limiting the mainstream cpu's to 4 cores with HT,i know that they have practically no competition from Amd apart from the integrated GPU
but i would like to see at least 6 core parts with HT,before upgrading again....
Just my thoughts.............
Bad temperatures are due to the contact between the IHS and die, not because of the TIM, but because of the black adhesive they used.
Senior Member
Posts: 6814
Joined: 2006-01-18
I will wait for officiall benchmarks first to see how good they will overclock assuming they will be using fluxless solder this time and not that cheap thermal compound like ivy,to be honest i personally think they might have intentionally used that cheap thermal paste on ivy to limit it's oc potential so more consumers coming let's say from Sandy bridge will be more tempted to upgrade to the new platform and not seeing it as a side grade, just imagine an Ivy bridge running at 5ghz on air performing similarly with a Haswell at let's say 4.6ghz (just an example) that for me is more of a side grade then an upgrade,and limiting the mainstream cpu's to 4 cores with HT,i know that they have practically no competition from Amd apart from the integrated GPU
but i would like to see at least 6 core parts with HT,before upgrading again....
Just my thoughts.............
Yes, because Intel sabotages all its parts for the 0.1% of people that overclock their CPU.
How did you ever figure out Intel's master plan?
Moderator
Posts: 29445
Joined: 2007-09-19
Isnt each generation easier to OC ?
Senior Member
Posts: 1209
Joined: 2008-10-08
nope.
On topic;
Im aiming for 5.2Ghz
125mhz x 42multi
I will wait for officiall benchmarks first to see how good they will overclock assuming they will be using fluxless solder this time and not that cheap thermal compound like ivy,to be honest i personally think they might have intentionally used that cheap thermal paste on ivy to limit it's oc potential so more consumers coming let's say from Sandy bridge will be more tempted to upgrade to the new platform and not seeing it as a side grade, just imagine an Ivy bridge running at 5ghz on air performing similarly with a Haswell at let's say 4.6ghz (just an example) that for me is more of a side grade then an upgrade,and limiting the mainstream cpu's to 4 cores with HT,i know that they have practically no competition from Amd apart from the integrated GPU
but i would like to see at least 6 core parts with HT,before upgrading again....
Just my thoughts.............