First 12-core VMmark for Istanbul AppearsJune 10, 2009
VMware has posted the VMmark score for the first Istanbul-based system and it’s from HP: the ProLiant DL385 G6. While it’s not at the top of the VMmark chart at 15.54@11 tiles (technically it is at the top of the 12-core benchmark list), it still shows a compelling price-performance picture.
Comparing Istanbul’s VMmark Scores
For comparison’s sake, we’ve chosen the HP DL385 G5 and HP DL380 G6 as they were configured for their VMmark tests. In the case of the ProLiant DL380 G6, we could only configure the X5560 and not the X5570 as tested so the price is actually LOWER on the DL380 G6 than the “as tested” configuration. Likewise, we chose the PC-6400 (DDR2/667, 8x8GB) memory for the DL 385 G5 versus the more expensive PC-5300 (533) memory as configured in 2008.
As configured for pricing, each system comes with processor, memory, 2-SATA drives and VMware Infrastructure Standard for 2-processors. Note that in testing, additional NIC’s, HBA, and storage are configured and such additions are not included herein. We have omitted these additional equipment features as they would be common to a deployment set and have no real influence on relative pricing.
Systems as Configured for Pricing Comparison
|HP ProLiant DL385 G5||Opteron 2384||2.7||8||8||64||667||$10,877.00|
|HP ProLiant DL385 G6||Opteron 2435||2.6||12||12||64||667||$11,378.00|
|HP ProLiant DL380 G6||Xeon X5560*||2.93||8||16||96||1066||$30,741.00|
Here’s some good news: 50% more cores for only 5% more (sound like an economic stimulus?) The comparison Nehalem-EP is nearly 3x the Istanbul system in price.
VMmark Price-Performance Comparison
|System||VMware Version||Score||Tiles||Ratio||Street||$/Tile||Workload Cost Factor|
|HP ProLiant DL385 G5||3.5U3||11.28||8||1.41||$10,877.00||$1,359.63||100.00%|
|HP ProLiant DL385 G6||4||15.54||11||1.41||$11,378.00||$1,034.36||76.08%|
|HP ProLiant DL380 G6||4||24.15||17||1.42||$30,741.00||$1,808.29||133.00%|
For VMmark comparisons – especially within the same processor family – we take the ratio of the VMmark score versus the number of tiles and compare it against “higher” scores in the same family to see how well the processors “scale.” In this case, the 2.6GHz Istanbul scales linearly maintaining a per-tile ratio of 1.41 – just like its slightly higher clocked sibling at 2.7GHz. Compared to Nehalem-EP, the Istanbul shows a near 2x cost-per-workload advantage within the same performance-per-tile ratio.
We expect the Istanbul to be capable of a higher VMmark score with the addition of more memory. At 80GB- DDR-667 (8x8GB + 4x4GB, 12 of 16 slots used), we estimate the Istanbul would be capable of managing a score of 16.9@12 tiles. At $100/stick, that would result in the following (estimated) VMmark and price-performance results:
|System (80GB/DDR2-667)||VMware Version||Score (est.)||Tiles (est.)||Ratio (est.)||Street (estimated)||$/Tile (estimated)||Workload Cost Factor|
|HP ProLiant DL385 G6||4||16.94||12||1.41||$11,778.00||$981.50||72.19%|
The AMD Opteron known as “Istanbul” has delivered on its price-performance promise. While it’s not king of the absolute performance hill, it does grab the virtualization price-performance crown. At nearly 2x the workload cost factor of Istanbul, Nehalem-EP has some catching-up to do. As we have suggested, we are approaching a time where the availability of memory is limiting the core/threading of today’s processors.
Nehalem-EP won’t compete with Istanbul in virtualization economically until memory prices catch-up to its threading capabilities. We predicted that EP’s biggest challenge was its apparantly short-sighted memory capacity (number of low-cost DIMMs supported). This prediction is playing-out as a key economic problem for EP in virtualization comparisons. For HPC and virtualization implementations where budget scaling is no object, Nehalem-EP (and similarly EX) has a good performance-performance argument.