SAP benchmarks for AMD Shanghai released
I just came across some certified SAP server benchmarks for the soon to be launched AMD 45nm Shanghai processor in comparison to some Intel Tigerton 65nm processor and Intel Dunnington 45nm processor.
| System | CPU | Cores | Sockets | Clock | Score |
| HP ProLiant DL785 G5 | Opteron 8384 | 4 | 8 | 2.7 GHz | 7010 |
| IBM System x3950 M2 | Xeon X7350 | 4 | 8 | 2.93 GHz | 6615 |
| IBM System x3950 M2 | Xeon X7460 | 6 | 8 | 2.66 GHz | 9200 |
While I’m not 100% certain that the Opteron 8384 is a Shanghai processor, the fact that it has 128 KB L1 cache and 512 KB L2 cache per core, and 6 MB L3 cache per processor is a pretty good indicator that it is Shanghai.
The Intel Dunnington server probably holds the edge because of its single die monolithic 6-core CPU whereas the Shanghai only has 4-cores.
Also, more Shanghai results here.
Categories: AMD, Benchmarks, Intel
Since they’ve officially launched them today, the Opteron benchmark site has been updated with results for SPECint_rate, SPECfp_rate and SPECjbb:
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_8796_8800,00.html
There are also VMmark results published for the Shanghai Opterons:
http://www.vmware.com/products/vmmark/results.html
AMD’s naturally comparing the results of the 2384/8384 (non-SE) with 80/90W TDP Xeons, which I’m sure you’ll find deceitful
.
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_8796_8800~128957,00.html
The SPECpower comparison they put up is deceitful because they compare their own best system with a mediocre Intel system that they themselves submitted and handpicked to lose. If this is the standard by which we go by, then anyone can pretty much spin a benchmark anyway they want since they can always find their competitor’s weaker system and compare it to their own best. AMD has made a mockery out of the entire benchmarking process.
You just "came across them" did you??
more like Intel PR kindly helped you find them…
at least be honest!
Again George how are they being deceitful. The full disclosure is available for everyone to see and they are NOT comparing their best system to the Intel as there is another submission with much better scores. You George are making a mockery of tech journalism.
It appears to me as if cherry-picking results isn’t a completely alien concept to you. I’m not surprised to find an 8P-comparison (2.7-hop topology for the Opterons with a diminishing scalability advantage) with Intels hex-core X7460 as a competitor in your blog. It would have surprised me on the other hand to have seen you reporting how the Shanghai-Opterons bury the competition in VMmark or SPECjbb2005.
Did you not see me praise the performance gains of AMD in the same piece criticize them for putting up suboptimal Intel benchmarks? Maybe if you spent less time apologizing for AMD, you would have noticed that.
I have not gotten around to analyzing all the new benchmarks yet because I have a different day job now. But when I do, I’ll be sure to praise AMD Shanghai for where it clearly is now the superior product (at least for the next 2 months until Nehalem arrives).