Decoupled L3 Cache

With Nehalem Intel introduced an on-die L3 cache behind a smaller, low latency private L2 cache. At the time, Intel maintained two separate clock domains for the CPU (core + uncore) and a third for what was, at the time, an off-die integrated graphics core. The core clock referred to the CPU cores, while the uncore clock controlled the speed of the L3 cache. Intel believed that its L3 cache wasn't incredibly latency sensitive and could run at a lower frequency and burn less power. Core CPU performance typically mattered more to most workloads than L3 cache performance, so Intel was ok with the tradeoff.

In Sandy Bridge, Intel revised its beliefs and moved to a single clock domain for the core and uncore, while keeping a separate clock for the now on-die processor graphics core. Intel now felt that race to sleep was a better philosophy for dealing with the L3 cache and it would rather keep things simple by running everything at the same frequency. Obviously there are performance benefits, but there was one major downside: with the CPU cores and L3 cache running in lockstep, there was concern over what would happen if the GPU ever needed to access the L3 cache while the CPU (and thus L3 cache) was in a low frequency state. The options were either to force the CPU and L3 cache into a higher frequency state together, or to keep the L3 cache at a low frequency even when it was in demand to prevent waking up the CPU cores. Ivy Bridge saw the addition of a small graphics L3 cache to mitigate this situation, but ultimately giving the on-die GPU independent access to the big, primary L3 cache without worrying about power concerns was a big issue for the design team.

When it came time to define Haswell, the engineers once again went to Nehalem's three clock domains. Ronak (Nehalem & Haswell architect, insanely smart guy) tells me that the switching between designs is simply a product of the team learning more about the architecture and understanding the best balance. I think it tells me that these guys are still human and don't always have the right answer for the long term without some trial and error.

The three clock domains in Haswell are roughly the same as what they were in Nehalem, they just all happen to be on the same die. The CPU cores all run at the same frequency, the on-die GPU runs at a separate frequency and now the L3 + ring bus are in their own independent frequency domain.

Now that CPU requests to L3 cache have to cross a frequency boundary there will be a latency impact to L3 cache accesses. Sandy Bridge had an amazingly fast L3 cache, Haswell's L3 accesses will be slower.

The benefit is obviously power. If the GPU needs to fire up the ring bus to give/get data, it no longer has to drive up the CPU core frequency as well. Furthermore, Haswell's power control unit can dynamically allocate budget between all areas of the chip when power limited.

Although L3 latency is up in Haswell, there's more access bandwidth offered to each slice of the L3 cache. There are now dedicated pipes for data and non-data accesses to the last level cache.

Haswell's memory controller is also improved, with better write throughput to DRAM. Intel has been quietly telling the memory makers to push for even higher DDR3 frequencies in anticipation of Haswell.

Feeding the Beast: 2x Cache Bandwidth in Haswell TSX
POST A COMMENT

248 Comments

View All Comments

  • Kepe - Friday, October 05, 2012 - link

    Let's see. I think we can agree that the Samsung Galaxy S III was the most important Android phone launch of the summer, so it should get comparable treatment if Anandtech was completely neutral. Let's compare the articles about the SGS III vs. iPhone 5

    Doing a search on anandtech.com gives us 8 articles/news posts about the SGS III vs. 13 articles/news posts about the iPhone 5.

    SGS 3:

    Five news stories about product announcements
    Performance Preview article
    Preview article
    Review article

    iPhone 5:

    Why iPhone 5 isn't launched in 2011 article
    Analyzing rumours about iPhone 5 article
    New SoC in iPhone 5 article
    iPhone 5 Live Blog from the product launch seremony
    Three news stories about new features and product announcements
    iPhone 5 Hands On article
    Lack of simultaneous voice and LTE/EVDO article
    Analyzing Geekbench results article
    Sunspider Performance Analysis article
    Performance Preview article
    iPhone 5 Display Thoroughly Analyzed article

    + The upcoming iPhone 5 Review article
    + articles such as "iOS6 Maps Thoroughly Investigated"

    Look at the difference. It's quite clear which device gets more coverage. And it's the same thing for older iPhones. Articles such as "Camping out for the new iPhone 3GS".

    This is NOT equal treatment of all products. This is why my trust for Anandtech has started to slip. Yes, Anandtech still is the best place for reviews, but one really has to wonder if those reviews still are as neutral and objective as they used to be.
    Reply
  • vFunct - Saturday, October 06, 2012 - link

    It's an android device. Android devices do not matter. Everyone uses iPhones anyways, They are better. Apple makes better products, including laptops.

    No need to waste space on Android.
    Reply
  • Haugenshero - Saturday, October 06, 2012 - link

    Please take your pointless apple fanboy drivel to another site that doesn't care about actual hardware and software and just like shiny things. Reply
  • cjl - Saturday, October 06, 2012 - link

    Apple's (iOS) current sales are only 20% of the overall smartphone market share, while Android is over 60%, so if either one of the two is largely irrelevant, it's apple. Reply
  • HisDivineOrder - Sunday, October 07, 2012 - link

    Shhhh, icks-nay on the facts-nay.

    You might cause a fanboy's head to explode near one of those inconveniently placed explosive barrels we walk by in real life.

    Just imagine a chain reaction. Caused by an Apple fan's mind being blown. You might take out an entire city block.

    Do you want that kind of devastation on your karma? Think different. ;)
    Reply
  • vFunct - Sunday, October 07, 2012 - link

    The Android market is the cheap giveaways.

    No one willingly pays money for an Android phone.

    Not everyone can afford the premium quality of an Apple product. They will have to settle for an inferior Android devices instead until they can afford higher quality products.
    Reply
  • Kepe - Monday, October 08, 2012 - link

    Nice trolling there. Now go back under the bridge and stay there =) Reply
  • Old_Fogie_Late_Bloomer - Tuesday, October 09, 2012 - link

    "Not everyone can afford the premium quality of an Apple product. They will have to settle for an inferior Android devices instead until they can afford higher quality products."

    Ha ha ha! Well done, if you're screwing around.

    But seriously, if you actually believe that, seek psychiatric help. :-P
    Reply
  • Spunjji - Thursday, October 18, 2012 - link

    Hahahahahahahahahahahahahahahahahahahahaha Reply
  • solipsism - Tuesday, October 09, 2012 - link

    And how many of the 20% is on one phone? Let me know when you figure out how to cover every single Androd-based device that hits the market in a given year. Reply

Log in

Don't have an account? Sign up now