mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > GMP-ECM

Reply
 
Thread Tools
Old 2012-04-05, 13:55   #1
R.D. Silverman
 
R.D. Silverman's Avatar
 
Nov 2003

1D2416 Posts
Default Intel Benchmarks

Does anyone have an Ivy Bridge benchmark for GMP-ECM?

I am skeptical (for several reasons) about improved performance
over (say) Sandy Bridge.
R.D. Silverman is offline   Reply With Quote
Old 2012-04-05, 14:02   #2
BigBrother
 
Feb 2005
The Netherlands

2·109 Posts
Default

Quote:
Originally Posted by R.D. Silverman View Post
Does anyone have an Ivy Bridge benchmark for GMP-ECM?

I am skeptical (for several reasons) about improved performance
over (say) Sandy Bridge.
They won't be officially released until the end of April. Maybe someone here has bought an early production sample on Ebay?

What makes you sceptical? The main improvements seem to be less energy consumption and a more powerful built-in GPU, which don't help GMP-ECM much. At least there's nothing that could make performance worse :)
BigBrother is offline   Reply With Quote
Old 2012-04-05, 15:17   #3
R.D. Silverman
 
R.D. Silverman's Avatar
 
Nov 2003

22×5×373 Posts
Default

Quote:
Originally Posted by BigBrother View Post
What makes you sceptical? The main improvements seem to be less energy consumption and a more powerful built-in GPU,
You just answered your own question! I see very little improvement in
memory bandwidth, memory latency, L2/L3 cache size, or CPU performance.
R.D. Silverman is offline   Reply With Quote
Old 2012-04-06, 00:06   #4
only_human
 
only_human's Avatar
 
"Gang aft agley"
Sep 2002

2·1,877 Posts
Default

For completeness, another addition is the new random number generator code named "Bull Mountain"
only_human is offline   Reply With Quote
Old 2012-04-06, 14:27   #5
TheJudger
 
TheJudger's Avatar
 
"Oliver"
Mar 2005
Germany

2×3×5×37 Posts
Default

Quote:
Originally Posted by R.D. Silverman View Post
You just answered your own question! I see very little improvement in
memory bandwidth, memory latency, L2/L3 cache size, or CPU performance.
Well, socket 1155 Sandy Bridge officially supports up to DDR3-1333 (two memory channels) while there is a very high chance that socket 1155 Ivy Bridge supports up to DDR3-1600 (two memory channels, too). So there is an improvement in memory bandwidth if you run those CPU at the specified speeds.

Oliver
TheJudger is offline   Reply With Quote
Old 2012-04-06, 14:29   #6
CRGreathouse
 
CRGreathouse's Avatar
 
Aug 2006

5,939 Posts
Default

Quote:
Originally Posted by only_human View Post
For completeness, another addition is the new random number generator code named "Bull Mountain"
This is one of the most exciting pieces for me! I'm not sure I like the implementation, but the idea of real hardware random number generation is pretty cool.
CRGreathouse is offline   Reply With Quote
Old 2012-04-11, 07:20   #7
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

22×2,239 Posts
Default

Ten thousand years of i7-3960X!!!!
(a SB, I don't own one yet, but if IB will be proved to be a "kepler" of the CPU, then I will prefer the "Fermi" SB).
LaurV is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
LLR Benchmarks pinhodecarlos Software 22 2017-04-28 19:13
Benchmarks davieddy Information & Answers 21 2013-04-28 11:16
Benchmarks ET_ Operazione Doppi Mersennes 18 2013-04-24 06:38
Where are the Benchmarks Sandman192 Homework Help 17 2012-04-05 19:03
Benchmarks Vandy Hardware 6 2002-10-28 13:45

All times are UTC. The time now is 07:29.

Wed Dec 2 07:29:45 UTC 2020 up 83 days, 4:40, 1 user, load averages: 1.38, 1.32, 1.34

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.