Intel's Haswell Architecture Analyzed: Building a New PC and a New Intel
by Anand Lal Shimpi on October 5, 2012 2:45 AM ESTPrioritizing ILP
Intel has held the single threaded performance crown for years now, but the why is really quite easy to understand: it has prioritized extracting instruction level parallelism with every generation. Couple that with the fact that every two years we see a "new" microprocessor architecture from Intel and there's a recipe for some good old evolutionary gains. The table below shows the increase in size of some major data structures inside Intel's architectures for every tock since Conroe:
| Intel Core Architecture Buffer Sizes | ||||||
| Conroe | Nehalem | Sandy Bridge | Haswell | |||
| Out-of-order Window | 96 | 128 | 168 | 192 | ||
| In-flight Loads | 32 | 48 | 64 | 72 | ||
| In-flight Stores | 20 | 32 | 36 | 42 | ||
| Scheduler Entries | 32 | 36 | 54 | 60 | ||
| Integer Register File | N/A | N/A | 160 | 168 | ||
| FP Register File | N/A | N/A | 144 | 168 | ||
| Allocation Queue | ? | 28/thread | 28/thread | 56 | ||
Increasing the OoO window allows the execution units to extract more parallelism and thus improve single threaded performance. Each generation Intel is simply dedicating additional transistors to increasing these structures and thus better feeding the beast.
This isn't rocket science, but it is enabled by Intel's clockwork fab execution. Designers can count on another 30% die area to work with every 2 years, so every 2 years they increase the size of these structures without worrying about ballooning the die. The beauty of evolutionary improvements like this is that when viewed over the long term they look downright revolutionary. Comparing Haswell to Conroe, the OoO scheduling window has grown by a factor of 2x, despite generation to generation gains of only 14 - 33%.
Facebook
Twitter
RSS
245 Comments
View All Comments
rundll - Friday, October 5, 2012 - link
Four cores and 95 W tdp.What is this?
meloz - Friday, October 5, 2012 - link
Yes this caught my eye and I would like an answer, too.Maybe it is one SKU with GT3 for desktop? Or maybe it is a 6 core part?
Or maybe.....it is the mother of all overclocking processors. Muhahahahah!
Kevin G - Friday, October 5, 2012 - link
I suspect that 95W is the rated socket limit. This is similar to how Intel advertises Ivy Bridge at 77 W on the desktop but tells motherboard manufacturers to build around the higher 95 W figure.What is odd is that Haswell will move some of the VRM circuitry on the package which should restrict just how far off that 95W figure motherboards can deviate.
meloz - Friday, October 5, 2012 - link
What a great article, Anand!Felt so good to read a 'proper' Anandtech article after so long, instead of the usual Apple worship and cheap fillers.
Haswell is looking very good. Would make an ideal upgrade for Sandy Bridge users. AMD is done, but thankfully Intel sees some threat from ARM so that will keep them innovating.
I hope Intel make a sensible choice with Haswell SKUs and get away from their artifical crippling and segmentation tendencies. That's about the only thing that can ruin Haswell.
Wolfpup - Friday, October 5, 2012 - link
Once again they bump up the number of transistors being used on their worthless video-and this time they even lower CPU performance (L3 cache) to appease their worthless video.Interesting article, but I guess I misunderstood previous articles...I thought Conroe through Ivy Bridge had 4 integer execution units per core? (As does Piledriver?)
haukionkannel - Friday, October 5, 2012 - link
Good article and information that you need win 8 to fully utilize Haswell was new information to me. It will be interesting to see how much better Haswell will be with win 8 compared to win 7. Seems to be same kind of dilemma as with AMD Bulldoser/piledriver where there seems to be some kind of better performance with new OS, but how much will reamain to be seen.Belard - Friday, October 5, 2012 - link
Apple owns various CPU tech and design companies such as P.A. Semi. They can build their own CPUs (not x86 of course)...Apple will do what they can to take out the middleman.
jwcalla - Friday, October 5, 2012 - link
Apple doesn't have any fabs though and if Samsung isn't willing to re-sign another contract, they're going to be in a bit of a bind. In other words, it won't be cheap. And even if Samsung does re-up, you can be sure that it'll come with an additional $1.05b price tag to offset any "losses" in their mobile division.I felt the first page overestimated Apple's influence quite a bit. They have ~5% desktop marketshare and 0% in the server space. Not to trivialize any loss in CPU sales, but Intel's primary headwinds don't involve a possible Apple switch to ARM.
Kevin G - Friday, October 5, 2012 - link
Apple's influence comes from the mobile market which is beginning to dwarf the PC market (and is larger than the server market in terms of volume). Apple is the largest tablet maker and a major smart phone manufacturer. There hardware is backed by one of the largest digital media markets. To do this Apple is the worlds largest consumer of flash memory whom orders are large enough to directly affect NAND pricing.With the rest of the industry going ultra mobile, they'll have to compete with Apple who is already entrenched. Sure the PC will survive but mainly for legacy work and applications. Their isn't enough of a PC market in the future to be viable long term with so many players.
jwcalla - Friday, October 5, 2012 - link
While all this is true, the first page seems to indicate that Intel is really pushing the low power envelop partly because of rumors that Apple will move away from Intel chips in their laptop / ultrabook products.While I'm sure Intel is happy to be in MBAs, etc., losing that business isn't going to be as big a deal as the other pressures facing the PC market (as you mention).
Now if WinRT on ultrabooks / laptops began to take off... that would be a huge problem for Intel.