AMD's 45nm Shanghai goes on sale

JPB

Diamond Member
Jul 4, 2005
4,064
89
91
Fud link


Several e-tailers have listed AMD's 45nm Shanghai server parts which are expected to officially debut in two weeks.

Branded Opteron 837X, 838X and 230X series, the new CPUs are priced from $814 for the cheapest of the lot, the 2380 clocked at 2.5GHz, while it's big brother, the 2382 at 2.6GHz costs $1,019.

Moving up, the 8378 at 2.4GHz costs $1,360, while the 8380 at 2.5GHz and 8382 at 2.6GHz are priced at $1,768 and $2,177 respectively.

The top range Opteron 8384 should end up clocked at 2.7GHz with a 75W TDP and it should sell for $2,509 or $2,240, depending on who you ask.
 

sunnn

Member
Oct 30, 2008
30
0
0
first post (just trying to pass time, 'been long time AT reader, 'thought i should contribute a little. also slightly amd biased- i have sempron and x1550 only slightly, i buy the cheapest and i love underdogs).

top nehalem is priced (rumored?) at $999 while top opteron is $2,509 or $2,240. hmm.
is this a sign of good things to come for amd?
 

extra

Golden Member
Dec 18, 1999
1,947
7
81
Or a sign of extremely wishful thinking. I just don't see how in the world that 2.7ghz opty is going to be competitive with the fastest clocked nehalem available, especially for over $2k per chip ? ? ? Am I missing something?
 

Yanagi

Golden Member
Jun 8, 2004
1,678
0
0
Nope, not a sign at all. You are comparing enterprise CPUs versus consumer CPUs. Apples to oranges im afraid.

Just to clarify:

The opteron is the Server version of the CPU.

Have a look at some Xeon MP Cpus, you'll find that they are also in the 1-2k price range.
The consumer edition of the cpu (will it be called phenom?) will have a much lower pricepoint,
 

Martimus

Diamond Member
Apr 24, 2007
4,488
153
106
Originally posted by: sunnn
first post (just trying to pass time, 'been long time AT reader, 'thought i should contribute a little. also slightly amd biased- i have sempron and x1550 only slightly, i buy the cheapest and i love underdogs).

top nehalem is priced (rumored?) at $999 while top opteron is $2,509 or $2,240. hmm.
is this a sign of good things to come for amd?

The top Nehalem price you quote is an expected desk top processor. Their Server based processors should be much more. They are expected to be inline with the Opteron prices (actually they are expected to be higher.) At least that is the way I understand it. (But who would listen to me, my daughter sure doesn't.)
 

Yanagi

Golden Member
Jun 8, 2004
1,678
0
0
Originally posted by: Martimus
Originally posted by: sunnn
first post (just trying to pass time, 'been long time AT reader, 'thought i should contribute a little. also slightly amd biased- i have sempron and x1550 only slightly, i buy the cheapest and i love underdogs).

top nehalem is priced (rumored?) at $999 while top opteron is $2,509 or $2,240. hmm.
is this a sign of good things to come for amd?

The top Nehalem price you quote is an expected desk top processor. Their Server based processors should be much more. They are expected to be inline with the Opteron prices (actually they are expected to be higher.) At least that is the way I understand it. (But who would listen to me, my daughter sure doesn't.)

Doesnt matter if its daughter, wife, sister or something else. Its in all womens nature not to listen to men
 

sunnn

Member
Oct 30, 2008
30
0
0
Originally posted by: Martimus
Originally posted by: sunnn
first post (just trying to pass time, 'been long time AT reader, 'thought i should contribute a little. also slightly amd biased- i have sempron and x1550 only slightly, i buy the cheapest and i love underdogs).

top nehalem is priced (rumored?) at $999 while top opteron is $2,509 or $2,240. hmm.
is this a sign of good things to come for amd?

The top Nehalem price you quote is an expected desk top processor. Their Server based processors should be much more. They are expected to be inline with the Opteron prices (actually they are expected to be higher.) At least that is the way I understand it. (But who would listen to me, my daughter sure doesn't.)

ok, ok i got it.
shanghai = server, deneb =desktop
nehalem = deskstop , ? = server
hmm, but shanghai is launching shortly while nehalem server is not, but the desktop one is. what im trying to say is, desktop and server chips almost have same cost to manufacture right (or am i wrong?), so as long as nehalem server is not available, shanghai and nehalem chips will be selling for the price quoted. means early on, amd will rip reward of their next gen chip while intel will sell nehalem for a desktop price until its server chip is ready (although the cost to manufacture is almost same).
i think its still a win for amd by having their server chip launching shortly instead launching desktop chip first.
 

soonerproud

Golden Member
Jun 30, 2007
1,874
0
0
Originally posted by: sunnn
Originally posted by: Martimus
Originally posted by: sunnn
first post (just trying to pass time, 'been long time AT reader, 'thought i should contribute a little. also slightly amd biased- i have sempron and x1550 only slightly, i buy the cheapest and i love underdogs).

top nehalem is priced (rumored?) at $999 while top opteron is $2,509 or $2,240. hmm.
is this a sign of good things to come for amd?

The top Nehalem price you quote is an expected desk top processor. Their Server based processors should be much more. They are expected to be inline with the Opteron prices (actually they are expected to be higher.) At least that is the way I understand it. (But who would listen to me, my daughter sure doesn't.)

ok, ok i got it.
shanghai = server, phenom =desktop
nehalem = deskstop , ? = server
hmm, but shanghai is launching shortly while nehalem server is not, but the desktop one is. what im trying to say is, desktop and server chips almost have same cost to manufacture right (or am i wrong?), so as long as nehalem server is not available, shanghai and nehalem chips will be selling for the price quoted. means early on, amd will rip reward of their next gen chip while intel will sell nehalem for a desktop price until its server chip is ready (although the cost to manufacture is almost same).
i think its still a win for amd by having their server chip launching shortly instead launching desktop chip first.

You got one thing wrong in your post. Phenom is the 65 nm desktop processor. Deneb will be the 45 NM desktop part.
 

Tanclearas

Senior member
May 10, 2002
345
0
71
Originally posted by: soonerproud
You got one thing wrong in your post. Phenom is the 65 nm desktop processor. Deneb will be the 45 NM desktop part.

Actually, 65 nm desktop processor (based on Barcelona) was Agena I believe. Deneb is the "code name", not the actual product name. The product name will probably still be Phenom, or perhaps end up with some distinguishing numbers/letters (like the Athlon XP/64/X2 models of the past).
 

coolpurplefan

Golden Member
Mar 2, 2006
1,243
0
0
Originally posted by: AlexWade
Bring on the reviews!

Yeah, I wish someone would buy one of them quick then post some comparison reviews with a similarly clocked Phenom. Then we might be able to estimate the rest.
 

soonerproud

Golden Member
Jun 30, 2007
1,874
0
0
Originally posted by: Tanclearas
Actually, 65 nm desktop processor (based on Barcelona) was Agena I believe. Deneb is the "code name", not the actual product name. The product name will probably still be Phenom, or perhaps end up with some distinguishing numbers/letters (like the Athlon XP/64/X2 models of the past).

I suspect AMD will be dropping the Phenom name for the Deneb processors for the same reason Microsoft is dropping the Vista naming scheme with Windows Seven. For too many people, Phenom is too much associated with over promising and under performing.

I stand corrected on the code names.
 

JaBro999

Member
Sep 14, 2006
93
0
0
Originally posted by: sunnn
ok, ok i got it.
shanghai = server, deneb =desktop
nehalem = deskstop , ? = server
hmm, but shanghai is launching shortly while nehalem server is not, but the desktop one is. what im trying to say is, desktop and server chips almost have same cost to manufacture right (or am i wrong?), so as long as nehalem server is not available, shanghai and nehalem chips will be selling for the price quoted. means early on, amd will rip reward of their next gen chip while intel will sell nehalem for a desktop price until its server chip is ready (although the cost to manufacture is almost same).
i think its still a win for amd by having their server chip launching shortly instead launching desktop chip first.


Server CPUs from Intel and AMD are based on the same architecture as their high volume, consumer/desktop CPU counterparts, but server CPUs are not always the same. Server CPUs may have RAS features lacking in desktop CPUs, support multiple socket systems (2, 4, 8, or more sockets), support different memory like FB-DIMMs, etc.

Also, high end server CPUs that go into multi-socket systems will have longer qualification periods than consumer/desktop oriented CPUs (Xeon MP, Itanium, etc. may need to be in partners' hands for a year of testing before servers runing the CPUs can go to market).

Intel has an additional challenge with the new Nehalem architecture because it is moving away from the P4 era FSB architecture to QuickPath. It is not necessarily surprising that Intel chose to focus Nehalem/QuickPath's debut on the high end consumer market, leaving more time for Intel and it's system partners (Dell, HP, IBM, etc.) to prepare for introducing the more complex, multi-socket server products.

The "delay" of Nehalem/QuickPath in the server space will give AMD some extra time to bring Shanghai to market, but AMD has its work cut out to prove that Shanghai can compete with Intel's current Xeon lineup (especially the 6 core Dunnington Xeons) let alone Intel's next generation CPUs.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: JaBro999
Originally posted by: sunnn
ok, ok i got it.
shanghai = server, deneb =desktop
nehalem = deskstop , ? = server
hmm, but shanghai is launching shortly while nehalem server is not, but the desktop one is. what im trying to say is, desktop and server chips almost have same cost to manufacture right (or am i wrong?), so as long as nehalem server is not available, shanghai and nehalem chips will be selling for the price quoted. means early on, amd will rip reward of their next gen chip while intel will sell nehalem for a desktop price until its server chip is ready (although the cost to manufacture is almost same).
i think its still a win for amd by having their server chip launching shortly instead launching desktop chip first.


Server CPUs from Intel and AMD are based on the same architecture as their high volume, consumer/desktop CPU counterparts, but server CPUs are not always the same. Server CPUs may have RAS features lacking in desktop CPUs, support multiple socket systems (2, 4, 8, or more sockets), support different memory like FB-DIMMs, etc.

Also, high end server CPUs that go into multi-socket systems will have longer qualification periods than consumer/desktop oriented CPUs (Xeon MP, Itanium, etc. may need to be in partners' hands for a year of testing before servers runing the CPUs can go to market).

Intel has an additional challenge with the new Nehalem architecture because it is moving away from the P4 era FSB architecture to QuickPath. It is not necessarily surprising that Intel chose to focus Nehalem/QuickPath's debut on the high end consumer market, leaving more time for Intel and it's system partners (Dell, HP, IBM, etc.) to prepare for introducing the more complex, multi-socket server products.

The "delay" of Nehalem/QuickPath in the server space will give AMD some extra time to bring Shanghai to market, but AMD has its work cut out to prove that Shanghai can compete with Intel's current Xeon lineup (especially the 6 core Dunnington Xeons) let alone Intel's next generation CPUs.

Exactly correct...
Another thing to remember is that the 8xxx Shanghai will work in 8 socket systems, while Nehalem won't have this ability until near the end of next year...
 

Cookie Monster

Diamond Member
May 7, 2005
5,161
32
86
Hm, but those 8 socket systems are still based on HT1.0, which could potentially hold back shanghai's performance by acting as a bottleneck. AMD too has to release these new platforms that support HT3.x+, and im sure this wont happen til maybe end of next year.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Cookie Monster
Hm, but those 8 socket systems are still based on HT1.0, which could potentially hold back shanghai's performance by acting as a bottleneck. AMD too has to release these new platforms that support HT3.x+, and im sure this wont happen til maybe end of next year.

The HT 3.0 platforms and chips are due late Q1 2009 at 45nm.
 

JaBro999

Member
Sep 14, 2006
93
0
0
Originally posted by: Viditor
Originally posted by: Cookie Monster
Hm, but those 8 socket systems are still based on HT1.0, which could potentially hold back shanghai's performance by acting as a bottleneck. AMD too has to release these new platforms that support HT3.x+, and im sure this wont happen til maybe end of next year.

The HT 3.0 platforms and chips are due late Q1 2009 at 45nm.


The current multi-processor Barcelona Opterons, and I assume this will be true of the first release of Shanghai, come with three HT links per CPU, so missing HT 3.0 on introduction won't be as bad as it might sound at first. The decision behind this is system was supposedly made by server vendors that wanted AMD to maintain socket/pin backwards compatibility with older Opteron systems and designs, so HT 3.0 had been put off until next year (that assumes there is also not a technical reason for the delay - who knows).

When AMD delivers multi-processor Opterons w/ HT 3.0, they should be getting one more HT link as well, for a total of four per CPU. With more HT links and more bandwidth per HT link, this will definitely help AMD in 4 and 8 way scaling with Shanghai. Let's hope so - the 4/8 way system space is the last area where AMD has an architectural advantage over Intel, and Intel's QuickPath promises to be at least as good if not better than HT.
 

Phynaz

Lifer
Mar 13, 2006
10,140
819
126
Originally posted by: Viditor
Originally posted by: JaBro999
Originally posted by: sunnn
ok, ok i got it.
shanghai = server, deneb =desktop
nehalem = deskstop , ? = server
hmm, but shanghai is launching shortly while nehalem server is not, but the desktop one is. what im trying to say is, desktop and server chips almost have same cost to manufacture right (or am i wrong?), so as long as nehalem server is not available, shanghai and nehalem chips will be selling for the price quoted. means early on, amd will rip reward of their next gen chip while intel will sell nehalem for a desktop price until its server chip is ready (although the cost to manufacture is almost same).
i think its still a win for amd by having their server chip launching shortly instead launching desktop chip first.


Server CPUs from Intel and AMD are based on the same architecture as their high volume, consumer/desktop CPU counterparts, but server CPUs are not always the same. Server CPUs may have RAS features lacking in desktop CPUs, support multiple socket systems (2, 4, 8, or more sockets), support different memory like FB-DIMMs, etc.

Also, high end server CPUs that go into multi-socket systems will have longer qualification periods than consumer/desktop oriented CPUs (Xeon MP, Itanium, etc. may need to be in partners' hands for a year of testing before servers runing the CPUs can go to market).

Intel has an additional challenge with the new Nehalem architecture because it is moving away from the P4 era FSB architecture to QuickPath. It is not necessarily surprising that Intel chose to focus Nehalem/QuickPath's debut on the high end consumer market, leaving more time for Intel and it's system partners (Dell, HP, IBM, etc.) to prepare for introducing the more complex, multi-socket server products.

The "delay" of Nehalem/QuickPath in the server space will give AMD some extra time to bring Shanghai to market, but AMD has its work cut out to prove that Shanghai can compete with Intel's current Xeon lineup (especially the 6 core Dunnington Xeons) let alone Intel's next generation CPUs.

Exactly correct...
Another thing to remember is that the 8xxx Shanghai will work in 8 socket systems, while Nehalem won't have this ability until near the end of next year...


Doesn't matter a bit, as Shanghai has to go up against Dunnington, which destroys it in an 8 socket system.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Phynaz



Doesn't matter a bit, as Shanghai has to go up against Dunnington, which destroys it in an 8 socket system.

And you know this because...

Just to be clear, there are NO benches of an 8 socket Shanghai system that I've heard of yet...do you actually know of any?
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Interesting subject. I would like to know . How Shanghai is suppose to compete with Nehalem . Viditor have you not seen the benchies that use many threads . Like what you would use in servers/ Nehalem performs just a bit less than 2 4core penryns.

I bet that 4core nehalem => than 2x 4core Shanghai. Read some of what Movieman at XT forums is doing with his setup.

Its nice AMD is getting 45nm. out in nov. But for servers Nehalem will completely render AMD servers useless. 4cores 8 threads. There is no rabbit in AMDs hat.

Also whats all this talk of Desktop only release of Nehalem . As far as I know DP server nehalem comes out at same time as XT nehalem . Or has that changed recentl?Or is it just fud started by someone?

http://www.crn.com/hardware/21...ChannelWebBreakingNews
 

JaBro999

Member
Sep 14, 2006
93
0
0
IBM's 8 way Dunnington benchmark submission for SAP-SD shows that Shanghai will have to be quite a performer to catch up:
http://www.sap.com/solutions/b...ark/sd2tier.epx?num=50

8 way Dunnington at 2.66 GHz (9/8/08): 9200 users / 46170 SAPS
System x3950 M2, Xeon MP X7460, 8s/48c, Win2K3 & DB2 9.5

8 way Barcelona at 2.5 GHz (10/22/08): 5800 users / 29670 SAPS
SF X4600M2, Opteron 8360 SE, 8s/32c, Solaris 10 & MaxDB 7.6:

I realize that is only one data point for one type of workload, but I'm skeptical that Shanghai will offer a large enough performance increase to close that 56% (or so) gap in performance with Dunnington. It seems that Shanghai really could use HT 3.0 and that extra link per core.
 

Idontcare

Elite Member
Oct 10, 1999
21,118
59
91
Yeah the way Shanghai performance was pitched as being 35% more (IPC and clockspeed increases combined) than Barcelona it doesn't look likely that it will compete with Dunnington in this app space.
 

Phynaz

Lifer
Mar 13, 2006
10,140
819
126
Originally posted by: Viditor
Originally posted by: Phynaz



Doesn't matter a bit, as Shanghai has to go up against Dunnington, which destroys it in an 8 socket system.

And you know this because...

Just to be clear, there are NO benches of an 8 socket Shanghai system that I've heard of yet...do you actually know of any?

Not any that I'm allowed to talk about yet

 

bryanW1995

Lifer
May 22, 2007
11,144
32
91
Originally posted by: Tanclearas
Originally posted by: soonerproud
You got one thing wrong in your post. Phenom is the 65 nm desktop processor. Deneb will be the 45 NM desktop part.

Actually, 65 nm desktop processor (based on Barcelona) was Agena I believe. Deneb is the "code name", not the actual product name. The product name will probably still be Phenom, or perhaps end up with some distinguishing numbers/letters (like the Athlon XP/64/X2 models of the past).

yeah, phenom was so wildly successful...they should keep the name around for a while longer. if they do decide to change it, I vote for "hector ruiz special".
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |