Intel's Haswell Architecture Analyzed: Building a New PC and a New Intel
by Anand Lal Shimpi on October 5, 2012 2:45 AM ESTThe Haswell Front End
Conroe was a very wide machine. It brought us the first 4-wide front end of any x86 micro-architecture, meaning it could fetch and decode up to 4 instructions in parallel. We've seen improvements to the front end since Conroe, but the overall machine width hasn't changed - even with Haswell.
Haswell leaves the overall pipeline untouched. It's still the same 14 - 19 stage pipeline that we saw with Sandy Bridge depending on whether or not the instruction is found in the uop cache (which happens around 80% of the time). L1/L2 cache latencies are unchanged as well. Since Nehalem, Intel's Core micro-architectures have supported execution of two instruction threads per core to improve execution hardware utilization. Haswell also supports 2-way SMT/Hyper Threading.
The front end remains 4-wide, although Haswell features a better branch predictor and hardware prefetcher so we'll see better efficiency. Since the pipeline depth hasn't increased but overall branch prediction accuracy is up we'll see a positive impact on overall IPC (instructions executed per clock). Haswell is also more aggressive on the speculative memory access side.
The image below is a crude representation I put together of the Haswell front end compared to the two previous tocks. If you click the buttons below you'll toggle between Haswell, Sandy Bridge and Nehalem diagrams, with major changes highlighted.
In short, there aren't many major, high-level changes to see here. Instructions are fetched at the top, sent through a bunch of steps before getting to the decoders where they're converted from macro-ops (x86 instructions) to an internally understood format known to Intel as micro-ops (or µops). The instruction fetcher can grab 4 - 5 x86 instructions at a time, and the decoders can output up to 4 micro-ops per clock.
Sandy Bridge introduced the 1.5K µop cache that caches decoded micro-ops. When future instruction fetch requests are made, if the instructions are contained within the µop cache everything north of the cache is powered down and the instructions are serviced from the µop cache. The decode stages are very power hungry so being able to skip them is a boon to power efficiency. There are also performance benefits as well. A hit in the µop cache reduces the effective integer pipeline to 14 stages, the same length as it was in Conroe in 2006. Haswell retains all of these benefits. Even the µop cache size remains unchanged at 1.5K micro-ops (approximately 6KB in size).
Although it's noted above as a new/changed block, the updated instruction decode queue (aka allocation queue) was actually one of the changes made to improve single threaded performance in Ivy Bridge.
The instruction decode queue (where instructions go after they've been decoded) is no longer statically partitioned between the two threads that each core can service.
The big changes in Haswell are at the back end of the pipeline, in the execution engine.
Facebook
Twitter
RSS
245 Comments
View All Comments
Magik_Breezy - Sunday, October 14, 2012 - link
Anything delivers "solid performance" on Facebook & iWorkWhy pay $2,000 for that?
random2 - Friday, October 5, 2012 - link
I agree. admittedly I am not an apple fan and view them as people who have undergone a degree of brainwashing compounded by the need for some to keep up with the Jone's. A certain degree of mind control must be necessary to stick with a company that has had some questionable business practices as far as customer relations, dealing with product issues and denying said issues, not to mention the whole hypocritical stance by apple in regards to copyright infringement has also left a bad taste in my mouth.hasseb64 - Saturday, October 6, 2012 - link
Disagree, not that much new from already published IDF reports almost 1 month ago. What is intresting is the claimed 40 EU GT3, other sources say lower amounts.JKflipflop98 - Saturday, October 6, 2012 - link
I totally agree. It's articles like this that have kept me coming back for years. Keep up the good work Anand!tipoo - Sunday, October 7, 2012 - link
"You can expect CPU performance to increase by around 5 - 15% at the same clock speed as Ivy Bridge. "That seems terribly disappointing for a tock, even IVB as a Tick managed 10% in most cases.
medi01 - Tuesday, October 9, 2012 - link
One can't be biased !@# !@#@ and a good journalist at the same time.One needs to be blind not to see how glass is always half empty for AMD, and half full for nVidia/Intel. F**!@#'s were shameless enough, to test 45W APU with 1000W PSU and such crap is all over the place.
Paulman - Friday, October 5, 2012 - link
As I was reading this article, about part way into the low platform power sections I suddenly had this thought: "Oh man, AMD is gonna die...!"I don't know if that's true for the entire microprocessor side of AMD, since they look like they're already starting to transition out of the desktop space, but I don't know if they're going to stand much of a chance if they're planning on entering the same TDP range as Haswell.
Do you think there's a chance AMD will start focussing on designing ARM ISA cores? Or will expanding on their x86 Bobcat-type cores be enough for them?
sean.crees - Friday, October 5, 2012 - link
I also worry about AMD. AMD has been 1-2 steps behind Intel for a while now, and now it seems Intel is at least 1 or 2 steps behind ARM and the future. Is that going to mean AMD is just too far behind to stay relevant now? If nothing else, i suppose AMD can fall back on graphic cards with it's ATI acquisition.Da W - Friday, October 5, 2012 - link
If Haswell keeps x86 relevant in the tablet space and thus Windows 8 has the upper edge over Windows RT and Windows tablets can grab +-50% market share from the iPad, then it can be good for AMD, provided they survive that long.RedemptionAD - Friday, October 5, 2012 - link
If AMD can create a team to focus on increasing IPC with a goal to one up Intel and have the ATI graphics people keep doing what they do with a time goal of say 2 years, (Note: Portables/Notebooks/Desktops should all be x64 by then), then I think that AMD will be able to return to their Athlon 64 glory days or better.