BM Seer Unofficial thoughts from an anonymous Sun employee

Sun UltraSPARC T2 & IBM Power6 comparison blogged about

Wednesday Aug 29, 2007

There is more preliminary UltraSPARC T2 performance is blogged about at: http://blogs.sun.com/jmeyer/entry/power6_goes_thud_part_v

Where John states:

    And IBM knows that next quarter, Sun will be introducing systems based on the new UltraSPARC T2, the world's first true system-on-a-chip and the world's fastest microprocessor. Preliminary estimates on one popular benchmark show that a single rack of UltraSPARC T2-based systems will outperform four racks of 4.7GHz POWER6-based p5 570s (more on that as we get closer to system announcement). No kidding.
I haven't seen this internal info yet, but I'll try to dig it up. Looking at other tests, I believe this one.

...John also talks more about the lagging IBM POWER6 rollout.

[2] Comments
Like this post? del.icio.us | furl | slashdot | technorati | digg
Posted at 01:35PM Aug 29, 2007 in Benchmark Issues
Tags:
Comments:

Okay, I've been a Sun fan for years. Great to see the capabilities of the T2. But I'd also love to know an apples-x2 comparison with the new QS21 Cell blades from IBM. For the SPARC64 VI, Sun quotes Linpack peak 1228.8 GFLOPS for an M9000 (which takes 3 42U racks?!). IBM's preliminary numbers for a full rack of QS21s (56 blades - I have no idea if this would require a supporting additional rack) is 25.8 TFLOPS. That's 20 times the M9000!

Now, I don't think those are official Linpack benchmarks for the QS21, and I wonder if they're quoting single-precision, since we all know the Cell has about 10x single- vs. double-precision performance. Are the M9000 numbers double-precision? Any similar numbers available yet for T2 systems?

Posted by Joe P. on August 31, 2007 at 09:52 AM PDT #

I don't know too much about the QS21s, IBM is being pretty vague about programming environment, delivered 64-bit vs. 32-bit perf, they only report peaks not linpack, likely only 32-bit peaks report now, each blade is limited to 2GB of memory,....

All M9000 is double precision and you can have TB of memory without distributed programming. Sun also have very dense blades with lots of memory for Xeon-based, Opteron-based, and UltraSPARC-based blades that can be mixed and matched in one chassis, etc.

I'm very cautious about IBM, I've gotten burnt a lot by them quoting peaks on special cases and not showing delivered stuff... sorry I'm cautious.

Posted by BM on August 31, 2007 at 11:58 AM PDT #

Post a Comment:
Comments are closed for this entry.