RWT Forums

By: EduardoS (no.delete@this.spam.com), July 19, 2012 6:19 pm

Kira (kirsc.delete@this.aeterna.ru) on July 19, 2012 5:44 pm wrote:
> Oracle's current roadmap says 3x throughput with no increase in
> single-thread performance. That seems like it would almost have to be either big
> strides in i/o (or memory) performance, or an increase in threads-per-core. The
> removal of references to better single-thread performance in recent roadmaps
> would indicate that a clock bump is probably not happening.

Interesting info, now let us think a bit about it, how to increase throughput by 3 times with either two times more cores or two times more threads per core without increasing single-thread performance?

< Previous Post in Thread

Next Post in Thread >

Topic	Posted By	Date
On the UltraSPARC T4	Gabriele Svelto	09/27/11 01:32 PM
"perceptron-based" based branch predictor	Mark Roulo	09/27/11 02:17 PM
On the UltraSPARC T4	DaveC	09/27/11 02:22 PM
On the UltraSPARC T4	Gabriele Svelto	09/27/11 03:28 PM
On the UltraSPARC T4	DaveC	09/27/11 03:59 PM
TPC-H benchmark result	Michael S	09/27/11 03:44 PM
TPC-H benchmark result	Jesper Frimann	09/28/11 01:43 AM
TPC-H benchmark result	Michael S	09/28/11 02:39 AM
TPC-H benchmark result	Jesper Frimann	09/28/11 03:49 AM
TPC-H benchmark result	Gabriele Svelto	09/28/11 09:44 AM
TPC-H benchmark result	Phil	09/28/11 05:56 AM
TPC-H benchmark result	Michael S	09/28/11 06:23 AM
TPC-H benchmark result	someone	09/28/11 06:54 AM
TPC-H benchmark result	Jesper Frimann	09/28/11 07:39 AM
cracking benchmarks	anon	09/28/11 02:10 PM
cracking benchmarks	Jesper Frimann	09/28/11 09:39 PM
TPC-H is not a good benchmark	David Kanter	09/28/11 05:02 PM
TPC-H benchmark result	someone	09/28/11 06:50 AM
TPC-H benchmark result	Michael S	09/28/11 07:11 AM
TPC-H benchmark result - the high priced spread	blaine	09/28/11 07:22 PM
TPC-H benchmark result - the high priced spread	Jesper Frimann	09/28/11 09:41 PM
TPC-H benchmark result - the high priced spread-SPARC T4 still comes out ahead!	Phil	09/29/11 12:06 AM
TPC-H benchmark result - the high priced spread-SPARC T4 still comes out ahead!	Jesper Frimann	09/29/11 01:49 AM
TPC-H benchmark result - the high priced spread-SPARC T4 still comes out ahead!	Phil	09/29/11 02:13 AM
TPC-H benchmark result - the high priced spread-SPARC T4 still comes out ahead!	Jesper Frimann	09/29/11 03:24 AM
TPC-H benchmark result	AC	10/10/11 01:11 AM
TPC-H benchmark result	Jesper Frimann	10/10/11 02:24 AM
But woulnd't that make TPC-H meaningless? (NT)	Daniel Bizo	10/11/11 12:29 AM
not very meaningful already	Michael S	10/11/11 01:09 AM
OLTP vs. analytic DBMS	David Kanter	10/11/11 02:44 PM
software or hardware	sp4mysf	09/29/11 06:14 AM
On the UltraSPARC T4	Thu Nguyen	09/27/11 04:40 PM
On the UltraSPARC T4	Brett	09/27/11 08:21 PM
On the UltraSPARC T4	David Kanter	09/27/11 08:33 PM
On the UltraSPARC T4	anon	09/27/11 08:56 PM
TERA MTA	Mark Roulo	09/28/11 08:36 AM
TERA MTA	anon	09/28/11 03:11 PM
TERA MTA	Michael S	09/29/11 12:58 PM
TERA MTA	anon	09/29/11 01:46 PM
TERA MTA	Brett	09/29/11 02:33 PM
Are you sure?	Michael S	09/29/11 03:19 PM
Are you sure?	Brett	09/29/11 05:31 PM
I trust The Register more	Paul A. Clayton	09/30/11 10:15 AM
TERA MTA	mpx	09/30/11 05:33 AM
On the UltraSPARC T4	anon	09/27/11 08:28 PM
On the UltraSPARC T4	Phil	09/28/11 12:23 AM
On the UltraSPARC T4	David Kanter	09/28/11 12:57 AM
On the UltraSPARC T4	Gabriele Svelto	09/28/11 09:22 AM
Database clustering	David Kanter	09/28/11 04:55 PM
Database clustering	Phil	09/29/11 12:15 AM
Database clustering	David Kanter	09/29/11 09:12 AM
Database clustering	Alex Fatkulin	09/29/11 11:14 AM
Database clustering	David Kanter	09/29/11 09:11 PM
Database clustering	Anon	09/30/11 12:32 AM
Database clustering	David Kanter	09/30/11 01:23 AM
Database clustering	mpx	10/05/11 01:12 AM
Database clustering	Jesper Frimann	09/30/11 02:14 AM
Database clustering	Anon	09/30/11 10:14 PM
Database clustering	Jesper Frimann	10/01/11 04:24 AM
Database clustering	Anon	10/01/11 05:27 PM
Real system level : )	David Kanter	10/02/11 06:59 PM
Real system level : )	Anon	10/05/11 11:36 PM
Database clustering	anon	10/02/11 05:34 PM
Database clustering	Jesper Frimann	10/03/11 01:56 AM
Database clustering	anon	10/03/11 08:55 AM
Database clustering	anon	10/03/11 11:41 AM
Database clustering	Jesper Frimann	10/03/11 02:24 PM
Google...	EduardoS	10/03/11 03:24 PM
Google...	Aaron Spink	10/04/11 01:02 AM
Database clustering	anon	10/03/11 04:04 PM
Database clustering	Jesper Frimann	10/04/11 02:42 AM
Database clustering	anon	10/04/11 08:13 AM
Database clustering	EduardoS	10/04/11 08:29 AM
Database clustering	anon	10/04/11 08:45 AM
Database clustering	bakaneko	10/04/11 09:11 AM
Database clustering	Gabriele Svelto	10/04/11 10:39 AM
Database clustering	anon	10/04/11 10:51 AM
Database clustering	Gabriele Svelto	10/04/11 01:27 PM
Database clustering	anon	10/04/11 01:39 PM
Database clustering	Jesper Frimann	10/04/11 11:26 AM
Database clustering	anon	10/04/11 01:37 PM
Database clustering	anonymous	10/02/11 10:27 PM
Database clustering	mpx	10/01/11 04:40 AM
Database clustering	Alex Fatkulin	09/30/11 05:42 AM
Database clustering	Jesper Frimann	09/30/11 06:19 AM
Database clustering	Alex Fatkulin	09/30/11 06:33 AM
Database clustering	Jesper Frimann	09/30/11 08:44 AM
Database clustering	Alex Fatkulin	09/30/11 09:01 AM
Database clustering	Jesper Frimann	09/30/11 09:55 AM
Database clustering	Alex Fatkulin	09/30/11 10:07 AM
Database clustering	Jesper Frimann	09/30/11 11:11 AM
Database clustering	Alex Fatkulin	09/30/11 11:33 AM
Database clustering	Jesper Frimann	09/30/11 03:00 PM
Database clustering	anon	10/01/11 09:18 AM
Database clustering	Michael S	10/01/11 10:33 AM
Database clustering	Jesper Frimann	10/01/11 10:22 PM
Database clustering	Jesper Frimann	10/12/11 11:14 PM
Warning: Bad pun (no content)	Paul A. Clayton	10/12/11 11:32 PM
Good one *8^) (NT)	Jesper Frimann	10/12/11 11:56 PM
Database clustering	Alex Fatkulin	10/04/11 08:09 AM
Database clustering	Jesper Frimann	10/04/11 11:35 AM
Database clustering	Gabriele Svelto	10/04/11 01:40 PM
Database clustering	David Kanter	10/04/11 02:45 PM
Database clustering	Jesper Frimann	10/05/11 12:09 AM
Database clustering	EduardoS	09/30/11 02:13 PM
Database clustering	Alex Fatkulin	10/04/11 05:56 AM
Database clustering	mpx	09/30/11 10:39 AM
Database clustering	EduardoS	09/30/11 02:07 PM
Database clustering	Alex Fatkulin	10/04/11 06:01 AM
Database clustering	mpx	10/05/11 01:02 AM
Database clustering	Jesper Frimann	10/01/11 01:32 AM
Database clustering	mpx	09/30/11 12:20 AM
Database clustering	EduardoS	09/30/11 02:09 PM
TPC-C king	Michael S	09/28/11 02:31 AM
On the UltraSPARC T4	someone	09/28/11 11:23 AM
Flash vs. disk on SPECjEnterprise2010?	Paul A. Clayton	09/27/11 03:29 PM
Flash vs. disk on SPECjEnterprise2010?	Gabriele Svelto	09/27/11 10:23 PM
Flash vs. disk on SPECjEnterprise2010?	Jesper Frimann	09/28/11 12:04 AM
Flash vs. disk on SPECjEnterprise2010?	Phil	09/28/11 12:32 AM
Are you trolling?	David Kanter	09/28/11 01:07 AM
memory bandwidth clarification etc..	Michael S	09/28/11 02:23 AM
memory bandwidth clarification etc..	Jesper Frimann	09/28/11 04:13 AM
memory bandwidth clarification etc..	Phil	09/28/11 06:06 AM
memory bandwidth clarification etc..	Jesper Frimann	09/28/11 09:25 AM
There is one (another one)...	EduardoS	09/28/11 02:54 AM
There is one (another one)...	Michael S	09/28/11 03:59 AM
Are you trolling? Yes, trolling for substantiation!!	Phil	09/28/11 05:47 AM
Are you trolling? Yes, trolling for substantiation!!	Jesper Frimann	09/28/11 11:01 AM
Are you trolling? Yes, trolling for substantiation!!	Phil	09/29/11 12:26 AM
Are you trolling? Yes, trolling for substantiation!!	Jesper Frimann	09/29/11 02:27 AM
Are you trolling? Yes, trolling for substantiation!!	Phil	09/29/11 04:46 AM
Are you trolling? Yes, trolling for substantiation!!	David Kanter	09/28/11 05:15 PM
Are you trolling? Yes, trolling for substantiation!!	Phil	09/29/11 12:53 AM
Are you trolling? Yes, trolling for substantiation!!	Jesper Frimann	09/29/11 07:43 AM
Pricing	David Kanter	09/28/11 05:28 PM
Pricing	Jesper Frimann	09/28/11 09:56 PM
Pricing	Michael S	09/29/11 01:42 PM
T4 better at cryptographic workloads?	Paul A. Clayton	09/28/11 05:49 AM
T4 better at cryptographic workloads?	Gabriele Svelto	09/28/11 09:36 AM
T4 better at cryptographic workloads?	Jesper Frimann	09/28/11 11:10 AM
T4 better at cryptographic workloads?	Gabriele Svelto	09/28/11 12:56 PM
T4 better at cryptographic workloads?	Max	09/29/11 06:53 AM
Instruction- vs. coprocessor-based extensions	Paul A. Clayton	09/28/11 04:48 PM
Memory versioning?	anon	10/02/11 08:48 PM
Memory versioning?	David Kanter	10/02/11 09:31 PM
Memory versioning--not just for TM?	Paul A. Clayton	10/09/11 02:37 PM
Memory versioning--not just for TM?	Gabriele Svelto	10/09/11 10:11 PM
Memory versioning--not just for TM?	David Kanter	10/11/11 03:15 PM
Point well taken	David Kanter	09/28/11 05:18 PM
T4 better at cryptographic workloads?	EduardoS	09/28/11 05:36 PM
Are you trolling?	Gabriele Svelto	09/28/11 09:49 AM
An addendum on I/O	David Kanter	09/28/11 11:29 AM
An addendum on I/O	Phil	01/27/12 08:19 AM
An addendum on I/O	anon	01/27/12 11:04 AM
An addendum on I/O	David Kanter	01/27/12 03:47 PM
An addendum on I/O	Kira	01/27/12 04:23 PM
Meant POWER7 (NT)	David Kanter	01/27/12 06:55 PM
An addendum on I/O	Jesper Frimann	01/29/12 07:58 AM
Are you trolling?	anon	09/28/11 01:03 PM
Flash vs. disk on SPECjEnterprise2010?	mpx	09/29/11 01:05 AM
Well done!	David Kanter	09/27/11 07:10 PM
Well done!	Gabriele Svelto	09/27/11 10:39 PM
On the UltraSPARC T4	mpx	09/29/11 12:56 AM
How much memory it supports?	EduardoS	10/11/11 01:37 AM
How much memory it supports?	Michael S	10/11/11 04:14 AM
Big power miss on T4?	David Kanter	07/16/12 09:24 PM
Big power miss on T4?	Gabriele Svelto	07/17/12 11:26 PM
Big power miss on T4?	David Kanter	07/19/12 04:04 PM
Big power miss on T4?	Kira	07/19/12 05:44 PM
Big power miss on T4?	EduardoS	07/19/12 06:19 PM
Big power miss on T4?	Kira	07/19/12 09:37 PM
Big power miss on T4?	David Kanter	07/19/12 10:15 PM
T4->T5	blaine	07/21/12 08:39 AM
Big power miss on T4?	EduardoS	07/20/12 08:18 AM
Big power miss on T4?	mpx	07/20/12 01:26 AM
Big power miss on T4?	EduardoS	07/20/12 08:26 AM
Big power miss on T4?	Gabriele Svelto	07/21/12 12:38 AM

Reply to this Topic
Name:
Email:
Topic:
Body:	No Text EduardoS (no.delete@this.spam.com) on July 19, 2012 6:19 pm wrote: > Kira (kirsc.delete@this.aeterna.ru) on July 19, 2012 5:44 pm wrote: > > Oracle's current roadmap says 3x throughput with no increase in > > single-thread performance. That seems like it would almost have to be either big > > strides in i/o (or memory) performance, or an increase in threads-per-core. The > > removal of references to better single-thread performance in recent roadmaps > > would indicate that a clock bump is probably not happening. > > Interesting info, now let us think a bit about it, how to increase throughput by 3 times with either two > times more cores or two times more threads per core <b>without</b> increasing single-thread performance? >
How do you spell blue?

RWT on Twitter