Article: Tukwila Update
By: someone (someone.delete@this.somewhere.com), February 7, 2009 9:10 am
Room: Moderated Discussions
Linus Torvalds (torvalds@linux-foundation.org) on 2/7/09 wrote:
---------------------------
[..]
>>1) IPF QPI and Xeon QPI don't use the same coherency
>>protocol (Xeon uses a simplified snoopy scheme suitable
>>for 2s/4s, IPF uses a more complex scalable distributed
>>directory system). Where do you think most of the "high
>>level" complexity is.
>
>If true (and I have no real reason to doubt it), that's
>just crazy, but I guess it meshes well with the whole
>insane notion of Intel just traditionally not letting x86
>grow up. I had really thought they got over that.
The scalability stuff in Tukwila takes up a significant
portion of the die area and power budget. Putting even
a fraction of that overhead in the same die that Intel
sells as desktop and laptop processors is just plain
nuts.
What about Beckton (Nehalem generation Xeon MP)?
Well the fact is x86 server sales of 4 sockets and up is
only worth about $1B per quarter, has flat sales growth
and AMD takes a bigger share of it than any other x86
segment. Putting the scalability stuff in Beckton would
probably cost it a couple of cores and maybe a speed
bin. Beckton will compete with future AMD processors
far more than IPF and extra cores and clock speed are
more valuable than scaling well to beyond 4 sockets
(which AMD server processors don't).
The truth is IPF is now more valuable to Intel than Xeon
MP. In the past five years IPF has severely curtailed x86's
presence in the mid and high end server segments. IPF
sales already exceed the total value of x86 server sales
of 4 sockets and up by about a third and is growing much
faster.
That's why IPF QPI gets distributed directory and x86 QPI
doesn't. Sorry, no craziness or conspiracy theory here.
It's just Intel doing what the server market wants. :-)
---------------------------
[..]
>>1) IPF QPI and Xeon QPI don't use the same coherency
>>protocol (Xeon uses a simplified snoopy scheme suitable
>>for 2s/4s, IPF uses a more complex scalable distributed
>>directory system). Where do you think most of the "high
>>level" complexity is.
>
>If true (and I have no real reason to doubt it), that's
>just crazy, but I guess it meshes well with the whole
>insane notion of Intel just traditionally not letting x86
>grow up. I had really thought they got over that.
The scalability stuff in Tukwila takes up a significant
portion of the die area and power budget. Putting even
a fraction of that overhead in the same die that Intel
sells as desktop and laptop processors is just plain
nuts.
What about Beckton (Nehalem generation Xeon MP)?
Well the fact is x86 server sales of 4 sockets and up is
only worth about $1B per quarter, has flat sales growth
and AMD takes a bigger share of it than any other x86
segment. Putting the scalability stuff in Beckton would
probably cost it a couple of cores and maybe a speed
bin. Beckton will compete with future AMD processors
far more than IPF and extra cores and clock speed are
more valuable than scaling well to beyond 4 sockets
(which AMD server processors don't).
The truth is IPF is now more valuable to Intel than Xeon
MP. In the past five years IPF has severely curtailed x86's
presence in the mid and high end server segments. IPF
sales already exceed the total value of x86 server sales
of 4 sockets and up by about a third and is growing much
faster.
That's why IPF QPI gets distributed directory and x86 QPI
doesn't. Sorry, no craziness or conspiracy theory here.
It's just Intel doing what the server market wants. :-)
Topic | Posted By | Date |
---|---|---|
Tukwila Update - article online | David Kanter | 2009/02/04 11:03 PM |
Tukwila Update - article online | Dan | 2009/02/05 02:17 AM |
Tukwila Update - article online | Joe Chang | 2009/02/05 08:16 AM |
Tukwila Update - article online | Temp | 2009/02/05 08:25 AM |
Tukwila Update - article online | Paul | 2009/02/05 11:29 AM |
Tukwila Update - article online | David Kanter | 2009/02/05 05:32 PM |
Tukwila Update - article online | Phil | 2009/02/06 12:24 AM |
Great. Finally hard numbers | Michael S | 2009/02/06 03:46 AM |
Tukwila Update - article online | lubemark | 2009/02/06 04:54 AM |
Tukwila Update - article online | Phil | 2009/02/06 06:29 AM |
Tukwila Update - article online | RagingDragon | 2009/02/07 02:39 PM |
Tukwila Update - article online | Michael S | 2009/02/07 03:09 PM |
Tukwila Update - article online | savantu | 2009/02/06 05:23 AM |
Tukwila Update - article online | Michael S | 2009/02/06 06:13 AM |
Tukwila Update - article online | someone | 2009/02/06 06:18 AM |
Tukwila Update - article online | Phil | 2009/02/06 06:47 AM |
Tukwila Update - article online | someone | 2009/02/06 07:17 AM |
Tukwila Update - article online | RagingDragon | 2009/02/07 02:51 PM |
Tukwila Update - article online | Linus Torvalds | 2009/02/06 07:37 AM |
Tukwila Update - article online | someone | 2009/02/06 08:19 AM |
Tukwila Update - article online | savantu | 2009/02/06 09:19 AM |
Tukwila Update - article online | Linus Torvalds | 2009/02/06 09:40 AM |
Tukwila Update - article online | savantu | 2009/02/06 10:00 AM |
Tukwila Update - article online | Phil | 2009/02/09 03:54 AM |
Tukwila Update - article online | Doug Siebert | 2009/02/09 09:40 PM |
Tukwila Update - article online | Jouni Osmala | 2009/02/10 12:03 AM |
Tukwila Update - article online | someone | 2009/02/10 05:15 AM |
Tukwila Update - article online | slacker | 2009/02/10 05:22 PM |
Tukwila Update - article online | Michael S | 2009/02/05 02:56 PM |
Tukwila Update - article online | David Kanter | 2009/02/05 03:55 PM |
Tukwila Update - article online | someone | 2009/02/05 04:47 PM |
Tukwila Update - article online | anon | 2009/02/05 09:16 PM |
Tukwila Update - article online | RagingDragon | 2009/02/05 09:27 PM |
Tukwila Update - article online | someone | 2009/02/06 06:32 AM |
Tukwila Update - article online | anon | 2009/02/06 08:25 AM |
Tukwila Update - article online | someone | 2009/02/06 08:40 AM |
POWER6 memory bandwidth | Michael S | 2009/02/05 02:30 AM |
POWER6 memory bandwidth | someone | 2009/02/05 06:00 AM |
POWER6 memory bandwidth | Michael S | 2009/02/05 06:36 AM |
POWER6 interconnect | confused | 2009/02/05 09:50 AM |
POWER6 interconnect | foobar | 2009/02/05 01:12 PM |
POWER6 memory bandwidth | Wes Felter | 2009/02/05 11:57 AM |
POWER6 memory bandwidth | Jesper Frimann | 2009/02/09 10:54 PM |
POWER6 memory bandwidth | Michael S | 2009/02/10 06:21 AM |
Why the platform focus? | Linus Torvalds | 2009/02/05 07:40 AM |
Why the platform focus? | savantu | 2009/02/05 07:50 AM |
Why the platform focus? | Vincent Diepeveen | 2009/02/05 08:29 AM |
Why the platform focus? | savantu | 2009/02/05 09:34 PM |
Why the platform focus? | Vincent Diepeveen | 2009/02/05 10:09 PM |
Why the platform focus? | Phil | 2009/02/06 12:10 AM |
Why the platform focus? | savantu | 2009/02/06 12:50 AM |
Why the platform focus? | Phil | 2009/02/06 06:09 AM |
Why the platform focus? | savantu | 2009/02/06 09:08 AM |
Why the platform focus? | someone | 2009/02/06 09:21 AM |
Why the platform focus? | mpx | 2009/02/06 01:04 PM |
Why the platform focus? | someone | 2009/02/06 01:16 PM |
Why the platform focus? | RagingDragon | 2009/02/07 03:16 PM |
Why the platform focus? | mas | 2009/02/25 07:28 AM |
itanium bigger than entire car industry | Vincent Diepeveen | 2009/02/06 06:12 AM |
itanium bigger than entire car industry | Devon Welles | 2009/02/06 06:51 AM |
itanium bigger than entire car industry | Vincent Diepeveen | 2009/02/06 09:41 AM |
itanium bigger than entire car industry | Dean Kent | 2009/02/06 06:56 PM |
Unit sales is meaningless when ASP grows faster | someone | 2009/02/07 08:38 AM |
Unit sales is meaningless when ASP grows faster | Dean Kent | 2009/02/07 02:10 PM |
Unit sales is meaningless when ASP grows faster | RagingDragon | 2009/02/07 03:34 PM |
itanium bigger than entire car industry | Vincent Diepeveen | 2009/02/08 04:35 AM |
itanium bigger than entire car industry | RagingDragon | 2009/02/07 03:40 PM |
Why the platform focus? | Vincent Diepeveen | 2009/02/06 06:47 AM |
Yes it doesm performance matters | bob | 2009/02/05 09:51 AM |
Yes it doesm performance matters | Venki | 2009/02/05 10:06 AM |
Why the platform focus? | Doug Siebert | 2009/02/06 12:07 AM |
Why the platform focus? | savantu | 2009/02/06 01:00 AM |
Why the platform focus? | someone | 2009/02/05 09:49 AM |
Why the platform focus? | Linus Torvalds | 2009/02/05 11:03 AM |
Why the platform focus? | Default | 2009/02/05 12:29 PM |
Why the platform focus? | anon | 2009/02/05 01:08 PM |
Why the platform focus? | someone | 2009/02/05 01:24 PM |
Why the platform focus? | Linus Torvalds | 2009/02/05 02:30 PM |
Why the platform focus? | Paradox | 2009/02/05 10:22 AM |
Why the platform focus? | slacker | 2009/02/05 12:41 PM |
Why the platform focus? | RagingDragon | 2009/02/05 09:57 PM |
Why the platform focus? | Michael S | 2009/02/06 05:11 AM |
Why the platform focus? | slacker | 2009/02/06 12:58 PM |
Why the platform focus? | Michael S | 2009/02/08 01:24 AM |
Why the platform focus? | someone | 2009/02/08 08:38 AM |
Why the platform focus? | David Kanter | 2009/02/08 03:27 PM |
Why the platform focus? | someone | 2009/02/08 06:26 PM |
Why the platform focus? | savantu | 2009/02/08 11:35 PM |
Why the platform focus? | someone | 2009/02/08 08:53 AM |
All x86 SpecInt scores are useless due to autopar (NT) | Michael S | 2009/02/05 02:15 PM |
Auto parallelization | David Kanter | 2009/02/05 05:17 PM |
All x86 SpecInt scores are useless due to autopar (NT) | Paradox | 2009/02/06 07:47 AM |
Why the platform focus? | David Kanter | 2009/02/05 03:49 PM |
Why the platform focus? | David Kanter | 2009/02/06 12:09 AM |
Why the platform focus? | Linus Torvalds | 2009/02/06 07:14 AM |
Why the platform focus? | savantu | 2009/02/06 09:37 AM |
Why the platform focus? | Linus Torvalds | 2009/02/06 11:49 AM |
Why the platform focus? | Linus Torvalds | 2009/02/06 12:09 PM |
Intel puts its money where its mouth is | someone | 2009/02/06 01:08 PM |
Intel puts its money where its mouth is | RagingDragon | 2009/02/07 05:01 PM |
Intel puts its money where its mouth is | someone | 2009/02/08 01:24 PM |
mission-critical | Michael S | 2009/02/08 04:06 PM |
mission-critical | mpx | 2009/02/09 01:30 AM |
mission-critical | rwessel | 2009/02/09 02:23 PM |
mission-critical | anon | 2009/02/09 02:55 AM |
mission-critical | EduardoS | 2009/02/09 04:17 PM |
mission-critical | Dean Kent | 2009/02/09 07:11 PM |
mission-critical | Michael S | 2009/02/10 04:20 AM |
mission-critical | Dean Kent | 2009/02/10 06:26 AM |
mission-critical | Michael S | 2009/02/10 07:01 AM |
mission-critical | Dean Kent | 2009/02/10 12:36 PM |
mission-critical | someone | 2009/02/10 08:05 AM |
mission-critical | Dean Kent | 2009/02/10 12:22 PM |
mission-critical | Zt | 2009/02/22 03:54 PM |
mission-critical | anon | 2009/02/10 09:41 PM |
mission-critical | EduardoS | 2009/02/10 12:46 PM |
mission-critical | Dean Kent | 2009/02/10 01:31 PM |
mission-critical | slacker | 2009/02/10 06:30 PM |
mission-critical | Vincent Diepeveen | 2009/02/18 06:20 AM |
Mission critical | mpx | 2009/02/09 12:00 AM |
Why the platform focus? | savantu | 2009/02/07 12:15 AM |
Sun and x86 server differentiation | David Kanter | 2009/02/07 12:34 AM |
Sun and x86 server differentiation | max | 2009/02/07 02:30 AM |
Sun and x86 server differentiation | someone | 2009/02/07 09:19 AM |
Sun and x86 server differentiation | Linus Torvalds | 2009/02/07 09:44 AM |
Sun and x86 server differentiation | RagingDragon | 2009/02/07 05:09 PM |
Sun and x86 server differentiation | Michael S | 2009/02/08 04:05 AM |
Sun and x86 server differentiation | RagingDragon | 2009/02/09 11:03 PM |
Sun and x86 server differentiation | Jesper Frimann | 2009/02/09 11:51 PM |
Sun and x86 server differentiation | Alex Jones | 2009/02/10 12:43 PM |
Why the platform focus? | bob | 2009/02/08 03:51 AM |
Why the platform focus? | someone | 2009/02/08 08:23 AM |
missing the big picture | AM | 2009/02/18 05:43 AM |
missing the big picture | Michael S | 2009/02/18 07:42 AM |
missing the big picture | AM | 2009/02/18 08:03 AM |
Why the platform focus? | mpx | 2009/02/06 11:47 AM |
Itanium - slowest and most obsolete server CPU family in the world, NOW. | mpx | 2009/02/06 03:48 PM |
Itanium - slowest and most obsolete server CPU family in the world, NOW. | Paul | 2009/02/07 01:56 PM |
z series? | Michael S | 2009/02/07 02:12 PM |
Itanium - slowest and most obsolete server CPU family in the world, NOW. | someone else | 2009/02/24 03:37 AM |
Itanium - slowest and most obsolete server CPU family in the world, NOW. | EduardoS | 2009/02/24 05:55 AM |
Itanium - slowest and most obsolete server CPU family in the world, NOW. | someone else | 2009/02/25 12:55 AM |
Itanium - slowest and most obsolete server CPU family in the world, NOW. | Michael S | 2009/02/25 01:27 AM |
Why the platform focus? | RagingDragon | 2009/02/07 05:18 PM |
Why the platform focus? | Paul | 2009/02/08 12:10 PM |
Why the platform focus? | Jukka Larja | 2009/02/08 10:04 PM |
Why the platform focus? | slacker | 2009/02/06 01:10 PM |
Why the platform focus? | Linus Torvalds | 2009/02/06 01:40 PM |
Why the platform focus? | savantu | 2009/02/06 01:51 PM |
Why the platform focus? | someone | 2009/02/06 01:58 PM |
Why the platform focus? | Linus Torvalds | 2009/02/07 08:26 AM |
Why the platform focus? | someone | 2009/02/07 09:10 AM |
Why the platform focus? | Linus Torvalds | 2009/02/07 09:40 AM |
Why the platform focus? | someone | 2009/02/07 11:24 AM |
Why the platform focus? | Doug Siebert | 2009/02/07 11:32 PM |
Why the platform focus? | max | 2009/02/08 03:57 AM |
Why the platform focus? | Michael S | 2009/02/08 04:20 AM |
Why the platform focus? | someone | 2009/02/08 08:15 AM |
Why the platform focus? | Doug Siebert | 2009/02/08 10:36 PM |
Why the platform focus? | hobold | 2009/02/09 04:49 AM |
Why the platform focus? | anon | 2009/02/24 12:57 AM |
Why the platform focus? | Linus Torvalds | 2009/02/24 08:45 AM |
Why the platform focus? | savantu | 2009/02/24 11:30 AM |
Why the platform focus? | slacker | 2009/02/24 12:51 PM |
Why the platform focus? | savantu | 2009/02/24 11:04 PM |
Why the platform focus? | Michael S | 2009/02/25 01:34 AM |
Why the platform focus? | anon | 2009/02/25 09:17 AM |
Why the platform focus? | max | 2009/02/25 10:15 AM |
Why the platform focus? | someone | 2009/02/24 04:43 PM |
Why the platform focus? | Doug Siebert | 2009/02/24 07:26 PM |
Why the platform focus? | Howard Chu | 2009/02/25 02:07 AM |
Why the platform focus? | someone | 2009/02/25 05:48 AM |
Why the platform focus? | someone | 2009/02/25 05:41 AM |
Why the platform focus? | Linus Torvalds | 2009/02/25 08:17 AM |
Why the platform focus? | someone | 2009/02/25 08:55 AM |
has anyone seen Tukwila silicon? | anon | 2009/02/25 09:38 AM |
Why the platform focus? | Linus Torvalds | 2009/02/25 10:05 AM |
Why the platform focus? | slacker | 2009/02/25 12:11 PM |
Why the platform focus? | a reader | 2009/02/26 08:11 PM |
Why the platform focus? | rcf | 2009/02/27 12:32 PM |
Why the platform focus? | max | 2009/02/27 01:11 PM |
Why the platform focus? | rcf | 2009/02/27 02:50 PM |
Why the platform focus? | Vincent Diepeveen | 2009/02/25 03:30 PM |
$40M sale to $16M company | bob | 2009/02/25 07:25 PM |
$40M sale to $16M company | Richard Cownie | 2009/02/26 11:21 AM |
Why the platform focus? | anonymous | 2009/02/24 10:52 AM |
Why the platform focus? | savantu | 2009/02/24 11:20 AM |
Why the platform focus? | anonymous | 2009/02/24 02:31 PM |
Why the platform focus? | savantu | 2009/02/24 11:05 PM |
Why the platform focus? | someone else | 2009/02/25 12:04 AM |
Why the platform focus? | Michael S | 2009/02/25 12:42 AM |
Put me down for $500 that Poulson doesn't arrive earlier than Q4/2011 (NT) | slacker | 2009/02/25 11:39 AM |
Why the platform focus? | someone | 2009/02/25 05:54 AM |
Why the platform focus? | anonymous | 2009/02/25 08:46 AM |
Why the platform focus? | someone | 2009/02/25 09:22 AM |
Why the platform focus? | anon | 2009/02/25 10:01 AM |
Why the platform focus? | anonymous | 2009/02/25 10:54 AM |
Why the platform focus? | mpx | 2009/02/24 01:11 PM |
Why the platform focus? | anon | 2009/02/24 07:57 PM |
Why the platform focus? | Doug Siebert | 2009/02/24 09:04 PM |
Why the platform focus? | anon | 2009/02/24 09:46 PM |
Why the platform focus? | Doug Siebert | 2009/02/25 04:13 PM |
Why the platform focus? | anon | 2009/02/25 07:53 PM |
Why the platform focus? | bob | 2009/02/25 08:00 PM |
Please try to keep up (NT) | anon | 2009/02/25 08:49 PM |
Why the platform focus? | Doug Siebert | 2009/02/25 11:09 PM |
Why the platform focus? | anon | 2009/02/26 12:12 AM |
Why the platform focus? | Michael S | 2009/02/26 01:16 AM |
Why the platform focus? | James | 2009/02/26 05:09 AM |
sufficiently intimate with the OS | Michael S | 2009/02/26 05:29 AM |
sufficiently intimate with the OS | anon | 2009/02/27 12:01 AM |
sufficiently intimate with the OS | Howard Chu | 2009/02/27 12:37 AM |
Why the platform focus? | Michael S | 2009/02/25 01:02 AM |
Why the platform focus? | anon | 2009/02/25 02:07 AM |
Why the platform focus? | anon | 2009/02/07 12:18 PM |
Why the platform focus? | Vincent Diepeveen | 2009/02/08 09:16 AM |
Why the platform focus? | anon | 2009/02/25 06:40 AM |
Intels financial status | Vincent Diepeveen | 2009/02/25 11:02 AM |
Why the platform focus? | someone | 2009/02/25 06:54 AM |
Why the platform focus? | Vincent Diepeveen | 2009/02/06 07:20 AM |
Why the platform focus? | Default | 2009/02/06 08:57 AM |
Why the platform focus? | Vincent Diepeveen | 2009/02/06 09:59 AM |
Why the platform focus? | RagingDragon | 2009/02/07 05:43 PM |
Tukwila Update - article online | Vincent Diepeveen | 2009/02/05 08:11 AM |