By: David Kanter (dkanter.delete@this.realworldtech.com), April 28, 2012 6:29 am
Room: Moderated Discussions
hcl64 (mario.smarq@gmail.com) on 4/27/12 wrote:
---------------------------
>David Kanter (dkanter@realworldtech.com) on 4/20/12 wrote:
>---------------------------
>>
>>You are correct that the caches are painfully slow, but that's not the reason why.
>>Frankly, I don't understand why the L1 is 4 cycles instead of 3. I REALLY don't
>>understand why the L2 cache is so slow (20 cycles, really??), because size alone
>>doesn't account for it. 12-14 cycles sounds much more reasonable.
>>
>>The L3 cache is also quite slow, in part because of the slow L2 and in part because
>>it runs at asynchronous to the cores. If you look at those two factors together
>>and assume a 14 cycle L2, you can probably cut the L3 latency down by ~10 cycles.
>>
>>
>>DK
>Good points. But isn't BD mostly asynchronous or semi->synchronous ?
No. You should read my article about Bulldozer.
http://www.realworldtech.com/page.cfm?ArticleID=RWT082610181333
Decoupled and asynchronous mean very different things. Asynchronous refers to frequency. Decoupled means there are buffers between stages. Those are two very different concepts and have very different implications for design.
DK
---------------------------
>David Kanter (dkanter@realworldtech.com) on 4/20/12 wrote:
>---------------------------
>>
>>You are correct that the caches are painfully slow, but that's not the reason why.
>>Frankly, I don't understand why the L1 is 4 cycles instead of 3. I REALLY don't
>>understand why the L2 cache is so slow (20 cycles, really??), because size alone
>>doesn't account for it. 12-14 cycles sounds much more reasonable.
>>
>>The L3 cache is also quite slow, in part because of the slow L2 and in part because
>>it runs at asynchronous to the cores. If you look at those two factors together
>>and assume a 14 cycle L2, you can probably cut the L3 latency down by ~10 cycles.
>>
>>
>>DK
>Good points. But isn't BD mostly asynchronous or semi->synchronous ?
No. You should read my article about Bulldozer.
http://www.realworldtech.com/page.cfm?ArticleID=RWT082610181333
Decoupled and asynchronous mean very different things. Asynchronous refers to frequency. Decoupled means there are buffers between stages. Those are two very different concepts and have very different implications for design.
DK



