32-bit vs. 64-bit Performance

Our entire benchmark suite to this point has been on 32-bit applications under a 32-bit OS, mostly because there are no good desktop 64-bit applications at this point in a popular 64-bit OS (not to mention the issues with 64-bit Windows XP we described earlier).

Under Linux however we don't have to wait for applications to be released in a 64-bit version, we can simply recompile them. Linux would thus provide us with an excellent venue to see the tangible performance increases from exposing the additional general purpose registers in 64-bit mode.

We ran all benchmarks on Red Hat Enterprise 2.9.5WS (Taroon), a beta release, booted in single user mode to avoid system services interfering with benchmark results. Neither Red Hat 9 nor 9.0.93 Beta (Severn) supply a 64-bit compiler or libraries, which is why we used Taroon.

The Taroon kernel initially had issues on startup requiring us to disable APIC and ACPI support to get it to install. Once actually running the OS was quite stable however DMA disk access was disabled for some reason.

We used the following compiler that came with Taroon:

gcc 3.2.3 20030502 (Red Hat Linux 3.2.3-16)

And the following kernel:

2.4.21-1.1931.2.393.ent

With this compiler and kernel we ran the following tests:

Whetstone

A simple C loop measuring floating point performance, configured to do double precision calculations.

Compiled with:
-O3 -msse2 -mfpmath=sse (and -m32 for 32bit, -m64 for 64bit)

The performance improvements due to 64-bit are in the 10 - 20% range we mentioned earlier.

Bytemark

An old integer CPU benchmark (FP results were discarded) - for more information on the tests visit this site.

Compiled with:
-O3 -msse2 -mfpmath=sse (and -m32 for 32bit, -m64 for 64bit)

Here we do see a small 2% drop in performance when moving to 64-bit in one test, however the rest of the tests show a 0 - 15% improvement across the board.

Lame 3.93

A MP3 encoder; encoded a 40minute .wav file (403MB).
Lame args: -b 192 -m s -h --quiet <file> - >/dev/null
(192kbps, simple stereo, high quality, output to nothing to avoid disk hits)

Compiled with:
-O3 -fomit-frame-pointer -fno-strength-reduce -malign-functions=4 -funroll-loops -ffast-math -msse2 -mfpmath=sse (again, -m32 for 32bit, -m64 for 64bit)

The performance improvement here is astounding - in 64-bit mode the Athlon 64 FX managed to finish the encode 34% quicker than in 32-bit mode, if these results are any hint of what could be in store for Windows users, there's a lot of promise behind the Athlon 64...assuming we get software support in time.

We wanted to do a transcode benchmark but that didn't work out - one library found a bug in gcc and transcode refused to compile. It actually forced a compile error because a structure came out padded, meaning they didn't expect anyone to run it on a 64bit machine just yet.

3D Rendering Final Words
Comments Locked

122 Comments

View All Comments

  • Anonymous User - Tuesday, September 23, 2003 - link

    #43 is a bit 486 DX style with 20 stages of pipeline up his crápperhole.
  • Anonymous User - Tuesday, September 23, 2003 - link

    ROFL@#36
    Dude the Athlon64 is a 32bit processor?....lol
    Hey everyone...the p4 is a 16bit cpu with 32bit extensions.

    Your an idiot #36. And this is coming from an Intel fanboy, so you really know your in the wrong.
  • Anonymous User - Tuesday, September 23, 2003 - link

    Price is more important for AMD because they've had their successes mainly on the price/performance front. If they are truly trying to match Intel on price, that advantage is essentially gone and it'll be an even harder battle to gain marketshare.

    Oh and some of you fanboys mustve missed the PM forum. What the industry sees is completely different from what fanboys see.
    http://www.anandtech.com/IT/showdoc.html?i=1873&am...
  • Anonymous User - Tuesday, September 23, 2003 - link

    Maybe if the intel & AMD both ran at the EXACT same clock it'd be fairer eh?

    I'd like to see more on the Opteron as I'm going to order 6 of them in November, thank the gods it's not my money.
  • Anonymous User - Tuesday, September 23, 2003 - link

    er... nm the above, I got it mixed up.
  • Anonymous User - Tuesday, September 23, 2003 - link

    FYI, in the LAME 3.93 MP3 encoding 32-bit vs 64-bit benchmark, you claim that 64-bit is 34% quicker when actually the graph shows it as 2/3.07*100 = 65% quicker.
  • dvinnen - Tuesday, September 23, 2003 - link

    Haha, this thread makes me laugh.

    1) The only thing Intel has that can toach FX-51 is the XeonMP, errrr, P4:EE. This processor doesn't start shipping for 2 to 3 months. I can understand including it in the review, but there should be some sort of disclaimer stating that this is a sample and may or not be reflective of the final product. The EE's also will only be released to the OEMs, so expect to have to pay VooDoo or AlienWare there outragous prices if you want one.

    2) The Intel fanboys ADMITT defeat. They are already rationlizing it by saying wait for Prescott to come out. All prescott is is a p4 clocked to 3.4 ghz with "improved hyperthreading." "The 11 new intruction sets" won't make any difference for a year or so (kind of like 64 bit goodness that you are bashing). But I guess the added bonus that you can heat a small house with it is something that AMD can't provide.

    And I wish people would stop complaing about the price. All new processors cost this much when they are first released. They'll come down, but not to the price of XPs for a while to come. The mobo costs will also drop drastcaly (past nForce2 prices?) over the comeing months beause of no north bridge and only a 6 layer PCB.
  • Anonymous User - Tuesday, September 23, 2003 - link

    64bit with 32bit compatibility would be what Itanium does. AMD64 is still native x86 with the ability to use 64-bit registers, thats why it can still run 32-bit programs as fast/faster than current CPUs.

    http://www.anandtech.com/cpu/showdoc.html?i=1815&a...
  • Anonymous User - Tuesday, September 23, 2003 - link

    #40 is a 64-bit moron.
  • Anonymous User - Tuesday, September 23, 2003 - link

    #36

    The Prescott is the next generation in the pentium family. It's not like it's a P4 with an increased multiplyer. AMD is in trouble when the Prescott comes rolling around.

Log in

Don't have an account? Sign up now