iXBT Labs - Computer Hardware in Detail






SPEC CPU2000. Part 8.
New versions of the Benchmark and Compilers

A month ago Intel announced new versions of its compilers and many are now interested in their relative performance. Besides, we were going to move to the new compiler Visual.NET from the obsolete MSVC 6 long ago.

The SPEC CPU2000 benchmark also changed its version number (though it was yet at the end of 2001 :) from 1.1 to 1.2. That is why today we will have a look at all these news and get prepared for work in 2003.

So, let me remind you what we already have:

  • Series iXBT.060202.*: MSVC 6 SP5 and Intel Compilers ver.5. 
  • Series iXBT.060602.*: MSVC 6 SP5 and Intel Compilers ver.6. 

Each series has versions regarding SIMD: without them, with MMX, with SSE and with SSE2. Each processor is usually tested on the maximum for it SIMD set. In particular, today all the tests used SSE2. 

Today we add the Intel Compiler 7.0 (both for C/C++ and for Fortran). Besides, together with the MSVC 6 SP5 we will try to use for libraries (the MSVC compiler is actually not used) a new studio version - Microsoft Visual.NET.

So, we have 6 combinations of compilers. All of them were tested on the following PC configuration:

  • Intel Pentium 4 2.53 GHz (FSB 533, L2 512K) 
  • ASUS P4T533 (i850E) mainboard
  • Samsung RDRAM PC4200 512MB memory 

Other components are not important for the SPEC CPU2000 tests. It can learn about it here.

The system ran under the Windows 2000 Pro SP3. First all the configurations were compiled (the OS was rebooted several times). After that the system was again reinstalled and the tests were started (note that the SPEC CPU2000 doesn't use any system files, that is why it's not necessary to reinstall the OS between separate tests).

On the whole, the scores are higher in the newer version, but inconsiderably:

Growth after the transition from MSVC6 SP5 to MS Visial.NET,%
  CINT2000 CFP2000
ic5 0.73 0.10
ic6 0.84 0.60
ic7 0.52 0.00


Relative performance of the Intel's compilers,%
CINT2000 msvc6 ms.net
ic5 100.00 100.00
ic6 99.90 100.00
ic7 100.42 100.21
CFP2000 msvc6 ms.net
ic5 100.00 100.00
ic6 102.13 102.63
ic7 103.95 103.85

These are the integral scores. If you look at the results of the separate tests, you will notice a greater difference. In particular, the transition to the Visual.NET comes to about 5 % in two subtests of CINT2000 and in CFP2000/183.equake the growth is over 13%. At the same time, for the CFP2000/188.ammp the scores fall down by almost 7%.

As to the new version 1.2, it supports several new operating systems and has corrections in the sources for better compatibility with various compilers. When we worked on this article, we had some problems with the benchmarks compilation for Linux which were resolved with the newer benchmark versions. So, this comparison is just formal (iXBT.060602.sse2.cfg configuration).

SPEC CPU2000 1.1 1.2
164.gzip 928 926
175.vpr 615 614
176.gcc 1159 1157
181.mcf 763 763
186.crafty 1006 1004
197.parser 922 925
252.eon 1097 1096
253.perlbmk 1163 1176
254.gap 1236 1232
255.vortex 1548 1548
256.bzip2 780 779
300.twolf 684 680
SPECint_base2000 960 959
SPEC CPU2000 1.1 1.2
168.wupwise 1276 1278
171.swim 1812 1811
172.mgrid 964 964
173.applu 1078 1078
177.mesa 989 985
178.galgel 1477 1477
179.art 742 742
183.equake 1099 1099
187.facerec 1232 1230
188.ammp 668 668
189.lucas 1407 1407
191.fma3d 976 976
200.sixtrack 470 469
301.apsi 758 758
SPECfp_base2000 1010 1010


So, we have two interesting results.

First of all, with the new compilers the integral scores grow up. If every per cent of performance is vital for you, you should certainly try the 7th version of the Intel's compiler together with the Microsoft Visual.NET. Note that the performance can increase by 5% and over on one application only. 

On the other hand, the difference is so slight that in future we will be able to use only new compilers and compare the scores with those obtained last year on the old configurations. It also refers to the SPEC CPU2000 version.

Kirill Kochetkov  (kochet@ixbt.com

Write a comment below. No registration needed!

Article navigation:

blog comments powered by Disqus

  Most Popular Reviews More    RSS  

AMD Phenom II X4 955, Phenom II X4 960T, Phenom II X6 1075T, and Intel Pentium G2120, Core i3-3220, Core i5-3330 Processors

Comparing old, cheap solutions from AMD with new, budget offerings from Intel.
February 1, 2013 · Processor Roundups

Inno3D GeForce GTX 670 iChill, Inno3D GeForce GTX 660 Ti Graphics Cards

A couple of mid-range adapters with original cooling systems.
January 30, 2013 · Video cards: NVIDIA GPUs

Creative Sound Blaster X-Fi Surround 5.1

An external X-Fi solution in tests.
September 9, 2008 · Sound Cards

AMD FX-8350 Processor

The first worthwhile Piledriver CPU.
September 11, 2012 · Processors: AMD

Consumed Power, Energy Consumption: Ivy Bridge vs. Sandy Bridge

Trying out the new method.
September 18, 2012 · Processors: Intel
  Latest Reviews More    RSS  

i3DSpeed, September 2013

Retested all graphics cards with the new drivers.
Oct 18, 2013 · 3Digests

i3DSpeed, August 2013

Added new benchmarks: BioShock Infinite and Metro: Last Light.
Sep 06, 2013 · 3Digests

i3DSpeed, July 2013

Added the test results of NVIDIA GeForce GTX 760 and AMD Radeon HD 7730.
Aug 05, 2013 · 3Digests

Gainward GeForce GTX 650 Ti BOOST 2GB Golden Sample Graphics Card

An excellent hybrid of GeForce GTX 650 Ti and GeForce GTX 660.
Jun 24, 2013 · Video cards: NVIDIA GPUs

i3DSpeed, May 2013

Added the test results of NVIDIA GeForce GTX 770/780.
Jun 03, 2013 · 3Digests
  Latest News More    RSS  

Platform  ·  Video  ·  Multimedia  ·  Mobile  ·  Other  ||  About us & Privacy policy  ·  Twitter  ·  Facebook

Copyright © Byrds Research & Publishing, Ltd., 1997–2011. All rights reserved.