Linux Shootout: Opteron 150 vs. Xeon 3.6 Nocona
by Kristopher Kubicki on August 12, 2004 2:35 PM EST- Posted in
- Linux
TSCP
We apologize for the broken TSCP Makefile in the previous review which rendered our initial results inaccurate. Fortunately we posted the file so that others were able to detect the error and not find fault with the processors instead. The large issue many of our readers have brought to our attention are the severe difference in performance between various optimizations. Below you can see how various compile flags affected our benchmark scores.
The first benchmark is run with the optimization flags:
-O2 -funroll-loops -frerun-cse-after-loop
The next benchmark is run with the optimization flags:
-O3 funroll-loops -frerun-cse-after-loop
Finally, we have the architecture optimized flags as well:
(Intel) -O3 - march=nocona -funroll-loops -frerun-cse-after-loop
(AMD) -O3 - march=k8 -funroll-loops -frerun-cse-after-loop
You are reading these charts correctly, the O3 flag actually penalizes the AMD CPU. We also compiled the program with -O2 -march=k8 but we got virtually the same score with or without the march flag.
We were informed others have been capable of much faster nodes per second using GCC 3.4.1 and the flagset:
-O3 -march=athlon-xp -funroll-loops -fomit-frame-pointer -ffast-math -fbranch-probabilities
We did not have time to fully test GCC 3.4.1, although there is a strong likelihood that 3.4 encourages better optimizations (particularly on the x86_64 platforms).
Crafty
For good measure, we have included Crafty into our chess benchmarks section. Crafty was only built using the "make linux-amd64" target. From the Makefile, it seems as though the "AMD64" moniker is slightly inappropriate. The target claims:
# -INLINE_AMD Compiles with the Intel assembly code for FirstOne(), # LastOne() and PopCnt() for the AMD opteron, only tested # with the 64-bit opteron GCC compiler.
The benchmark was generated by running the "bench" command inside the program.
It is clear the difference between both processors is quite severe in this instance. Although it is difficult to pin an exact culprit, there are likely multiple arch optimizations were left untapped, and thus our reasoning for discouraging overusage of optimizations in general.
Facebook
Twitter
RSS
92 Comments
View All Comments
tfranzese - Thursday, August 12, 2004 - link
Much better Kris, and I was only disappointed in not getting to see the 32-bit comparisons to see how well the current iAMD64 implementation is running.Only thing you left me confused on was the last paragraph on the John The Ripper benchmark. Might only be me who can't understand what you mean.
KristopherKubicki - Thursday, August 12, 2004 - link
datacipher:Datacipher, it really depends on your point of view. For server performance benchmarks, like the SQL stuff, another server chip was the way to go (like we did here).
If we go the desktop route, the line becomes a little blurry, particularly on Intel's end. Intel defines UP servers and high end workstations almost identically.
If we were to go on price alone, we still don't have the fairest comparison since the 150 is priced cheaper than the Xeon and the 3.6F. Some of my sources have said the 3.6F may debut considerably lower than its Xeon 3.6GHz server counterpart, even though they are the same processor with different pin outs.
Regardless of what you think, the 3.6F and Xeon 3.6GHz processors will compete against AMD dual channel offerings in the 2.4GHz range. If you read some of our other reviews, L2 cache size doesnot seem as critical on the A64 platform.
Whelp, anyway, hope that helps. When the 3.6F actually shows up at newegg with a price, then i will tell you for sure what it competes against :)
Kristopher
srg - Thursday, August 12, 2004 - link
Stellar Job on this one, you earnt the vacation. Anyway the reason this one seems more correct is that in the 32-bit tests, opteron beats the Xeon (so why should this one be so different), now everywhere else I've read are saying that the Xeon's 64-bit performance wasn't much and yours was a contradiction. That's why there was the contravercy. Anyway, this confirms the party line.srg
datacipher - Thursday, August 12, 2004 - link
Kristopher:Now I'm really confused! In the original article you say that the chips were not meant to compete...but then you also said things like "it turned out to be more appropriate than antipated". Now you are saying that since the 3.6F will be marketed against the 3500+...that it is the correct choice...
Now I'm trying to understand...forgive my ignorance. Basically you took a server chip as a stand in for the 3.6F against the 3500+ in what amounts to basically a desktop shootout? What I don't understand about this is how can Intel release a 3.6F with a roughly comparable cost to the 3500+ but with the same performance as the xeon 3.6?
Also, if the cpu choice was correct...why did you then switch to the opeteron 150?
Thanks in advance.
KristopherKubicki - Thursday, August 12, 2004 - link
Lynx516:I posted the GCC -v at the beginning of the review, please let me know if there is something else i should do.
Kristopher
KristopherKubicki - Thursday, August 12, 2004 - link
To be honest i wouldnt have known some of the mistakes i made had people not been so critical. I am not upset with the final outcome, it happens to everyone.Kristopher
Lynx516 - Thursday, August 12, 2004 - link
Sorry to be a bit harsh there Kris but they are not in the standard GCC3.3.3 manual which I was using as a reference. It would be nice to state that you are using a ported compiler in your config to prevent any future confusion. Though its a pretty good article over all. much better than some recent ones on Anand. I hope this standard is kept up.Humblest appologies
Lynx
Spectre999 - Thursday, August 12, 2004 - link
I posting because I was complaining about the first review on another board. The first review was horribly done.But it takes a real stand up person to take all the bitching that went on and be able to go back and look at what they did and decide to redo it because they made a mistake. For that I will give you much praise. The other hand is the first review shouldn't have been turned in the way it was but it isn't always the way someone does something that is the most important but the way they respond to the criticism they get.
So thanks for all the effort on redoing the article and you provided you are a stand up guy who simply made a mistake. It happens and everyone can move on.
manno - Thursday, August 12, 2004 - link
"Now will all of you A-Holes get off KrizK's & AT editorial staff's back!!"HAHHAHAHAHAHA I'm laughing my ass off.
Great Job getting in the first post, and a good first post at that.
-manno
hifisoftware - Thursday, August 12, 2004 - link
Good review. All sins are forgotten now :-)