Benchmarks show Opteron QC 1.9GHZ frags Clovertown 2.66GHZ
K10 1.9GHZ . Clovertown 2.66GHZ.
Intel has found it pretty hard to hide K10's massive 50% performance advantage. We expect Intel to immediately cut its prices by 50% or more, the 3GHZ Clovertown may be sold at $158, just below the 1.7GHZ Opteron QC.
Even Anand had to concede that K10 is fast. It is obvious that K10 has more than 20% integer IPC advantage over Intel.
posted by Sharikou, Ph. D. @ 9/09/2007 08:05:00 AM
16 Comments:
AMD Frags Honest Journalism:
http://www.bangkokpost.com/Database/05Sep2007_data006.php
Talk about abusive behavior. AMD fanboys need to learn that Intel making superior products is not "illegal" but trying to force journalists to literally lie in AMD's favor is beyond unethical.
http://www.xtremesystems.org/forums/showthread.php?t=158237
amd Opteron 2347 prefraged by Xeon L5320 *with the words of giant. ;) )
Sharikou created the 'pre-fragged' term ;-)
I believe it started when he made the ridiculous claim that K10 was so advanced that it has pre-fragged every upcoming architecture on Intel's roadmap.
Tick tock tick tock...INTEL CLOCKED
Time (Sec) in bold
IBM/IBM System X3455/Not applicable
AMD Barcelona 1.9 GHz 1 x 2 x 4 = 8
31526
3 Vehicle Collision
Intel/Mellanox Technolgies, Inc./Scali, Inc. Clovertown Blades/Infiniband DDR, OFED 1.2
Intel Xeon Clovertown 2.66GHz 1 x 2 x 4 = 8
32578
3 Vehicle Collision
Barcelona is too little too late when compared with Clovertown and Penryn. AMD has been pre-fragged. AMD has a habit of releasing new architectures that are pre-fragged by the competition! It happened in graphics now again in CPUs. This is an odd strategy AMD is pursuing.
AMD's CPUs are worthless. AMD is forced to sell top-end Barcelona CPUs at $1000. This is less than half of their old dual core pricing. AMD's pricing has collapsed, and ASP will fall rapidly. Expect growing losses. $600m this quarter and next at the minimum.
AMD is finished. BK in Q2'08.
Giant said...
Sharikou created the 'pre-fragged' term ;-)
I believe it started when he made the ridiculous claim that K10 was so advanced that it has pre-fragged every upcoming architecture on Intel's roadmap.
im sorry. :)
Even Phil Hester, AMD's CTO, admits that AMD is not a great company:
"The ATI acquisition itself is really about the long-term vision Hector had about what it takes to move AMD from a good company to a great company," Hester said
http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2007/09/09/BU2LS0RT2.DTL
Funny how most of the ex ATI mgmt has thrown in the towel.
Tick tock tick tock...AMD CLOCKED
Time (Sec) in bold
IBM/IBM System X3455/Not applicable
AMD Barcelona 1.9 GHz 1 x 2 x 4 = 8
49808
3 Vehicle Collision
Intel/Mellanox Technolgies, Inc./Scali, Inc. Clovertown Blades/Infiniband DDR, OFED 1.2
Intel Xeon Clovertown 2.66GHz 1 x 2 x 4 = 8
32578
3 Vehicle Collision
Ekekeke.. Intel is finished.. BK by Q2'08 for sure..
Pezal: You do realize you just called the AMD system slower right? That score is a TIME IN SECONDS (here's a hint: fewer seconds == faster time).
From your own post before you delete it to avoid embarrassment:
K10: 49808 (seconds)
Intel: 32578 (seconds)
I know you are also the drooling retard who could not even comprehend that a 2.4Ghz K10 will not actually be 485% faster than a 1.6Ghz K10, but please this is embarrassing.
Now onto something more intelligent, Sharikou, you and your multiple-personalities can tune out now, this involves logic, not racist mouth foaming.
So earlier I had posited that the K10 at 2Ghz would be in the range of a 2.4Ghz Intel quad core. It looks like I was wrong in 2 directions.
First, it appears that for most integer and 1-2 thread FP loads the K10 will actually be slower clock for clock than the current Intel Core 2's, with that gap widening somewhat further when Penryn comes out. For all you AMD fanboys screaming about SPECFP, note that it is actually the INTEL chips that are faster in the normal FP benchmarks, it is only in the SPECfp_rate that AMD excels (see below for more on that). Remember however, that the Intel still has a commanding lead at the SPECint_rate benchmark which is also measuring overall system throughput, and it favors Intel right now, meaning that the aging FSB is still able to handle most tasks that most servers in most markets need better than the K10 can.
In a few interesting (but also somewhat niche) areas, it appears K10 is actually faster than I originally predicted. AMD definitely has an advantage with its current memory controller in performance, and the specFP_rate shows that. You can read more about it right here. This will help AMD in the HPC sector, there is no denying that. However, in the vastly larger market out there for general-purpose servers, and for desktops and workstations (even ones doing heavy FP like rendering and games) the Intel chips are still noticeably faster. If you don't believe me... just look at AMD's prices for the K10... AMD is saying this with its own actions, not me.
The last thing to remember is this AMD: Sure you have the advantage now, but remember this:
1. No amount of screaming about "native" quadcore will make you profitable when your monstrously expensive dies cannot be sold for > $400, and even the artificially-inflated 8-way dies are only 1/2 of what you used to get for vastly cheaper Opterons just a year ago. Remember kids: The cost to make a 2Ghz K10 is EXACTLY the same as the cost to make a 3Ghz K10, but AMD can't do it, and if they could they'd be selling them, don't kid yourselves.
2. Yes you have an advantage in some HPC-type loads... but for the majority of FP loads used by most people every day, and for basically all non-FP loads the K10... your shining star that has cost the company billions... is slower than what Intel was selling at this time last year clock for clock.
As for the HPC segment, that's nice, but I hope to God AMD has something a hell of a lot better by next year since Intel's quickpath is going to erase any advantages AMD has in that niche right now.
pezal said...
Tick tock tick tock...AMD CLOCKED
Time (Sec) in bold
IBM/IBM System X3455/Not applicable
AMD Barcelona 1.9 GHz 1 x 2 x 4 = 8
49808
3 Vehicle Collision
Intel/Mellanox Technolgies, Inc./Scali, Inc. Clovertown Blades/Infiniband DDR, OFED 1.2
Intel Xeon Clovertown 2.66GHz 1 x 2 x 4 = 8
32578
3 Vehicle Collision
Ekekeke.. Intel is finished.. BK by Q2'08 for sure..
HAHAHAHAHAH what a douchebag kekekeke........further proof that all AMD fanbois are retards
'Twas the night before Barcelona, when all through the house
Not a creature was stirring, not even a Benchmark;
DUMBEST POST OF ALL TIME
Congratulations to Pezal!
Step 1: mi7chy posts fourth from the top the following accurate benchmark he found (I have no issue with mi7chy here) for "3 Vehicle Collision" from IBM.
Barcelona 31526
Xeon 32578
2. Pezal decides to take the Barcelona score and scales it up to a non-existent 3Ghz part
31526*(3/1.9)= 49808
3. Then he claims victory with his retarded "ekekeke".
PEZAL, the scores are in SECONDS. You are a LIAR and an IDIOT. Congrats on the DUMBEST fanboi post OF ALL TIME!
hahhahah i can just picture his dumbass sitting with a calculator extrapolating those results and typing them up like a retarded giddy school girl.
Little does he know after posting he will have claimed to the world that he is the biggest douchebag AMD fanboi of them all.
This kinda honor is only surpassed by sharikou and copyandpasteexpert......because of their lack of originality.....atleast pezal took his time and thought up that stupidity he posted.......kekekeke.....hahahhaha!!!
Benchmarks are flowing in,
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3092&p=5
K10 gives about ~15% IPC improvement over K8 clock for clock, per core.
What we know so far,
Conroe has about a ~25% performance advantage clock for clock over K8.
Still not enough to take down Intel's 65nm Kentsfield.
Since Penryn has about a 5% IPC improvement clock for clock over Kentsfield, AMD is going to have a hard time. Pricing is key for AMD.
This comment has been removed by the author.
Thanks for the link poke. 15% IPC improvement over K8 is not enough to compete with Intel. With Penryn Intel will enjoy higher clockspeeds and 10% IPC advantage.
AMD needs to compete with Intel on pricing, but with a massive die size this is just not happening.
AMD is finished. BK in Q2-08.
Post a Comment
<< Home