Linux and EM64T; Intel's 64-bit Suggestion
by Kristopher Kubicki on August 9, 2004 12:05 AM EST- Posted in
- Linux
Audio Encoding
Lame was compiled from source without optimizations. We only ran ./configure and make, without any flags. We realize that some people would like to verify our binaries and sample files for their own benchmarks. In order to save bandwidth and prevent copyright infractions, we will provide our test files and binaries under limited circumstances to serious inquiries. We ran lame on a 700MB .wav file using the command equivalent to the one below:# lame sample.wav -b 192 -m s -h >/dev/null
Encoding time, lower is better.
POV-RAY
Although POV-RAY is limited in application (particularly when compared against Mental Ray), it does provide a free open source solution for basic rendering. POV-Ray 3.50c was our choice of render engine for this benchmark. For benchmark specifics, we run the exact benchmark as specified by the POV-Ray official site. We use the precompiled RPM for this test.Render Time in Seconds, less is better.
POV-Ray does not have multithread support, so we were not surprised to see the HyperThreading configuration slowing down to the configuration without HT. We see the Athlon 64 processor pull way ahead; render tasks are extremely CPU and memory dependant. With the memory controller on the CPU, Athlon 64 becomes the stronger offering in this situation.
GZip
To throw in some rudimentary tests for GZip, we used the included GZip 1.3.5 to compress the .wav file from the benchmark above. We do not want to limit our I/O on writing to the hard drive, so the operation is performed as below:# time gzip -c sample.wav > /dev/null
Intel wins their first bout of the analysis, albeit not by much. We will find a recurring pattern later on with integer based calculations and the Nocona Xeon processor. The entire Prescott family of Intel CPUs received a dedicated integer multiplier rather than continually using the floating point multiplier. This becomes extremely useful in some of our other benchmarks.
Database Performance
We will run the standard SQL-bench suite included with RPM MySQL 4.0.20d.
Of all our benchmarks, the SQL-bench becomes the most baffling. The extremely threaded database application performs particularly poorly with HyperThreading enabled. The Althon 64 outperforms Intel again in this benchmark, and a lot of it is almost certainly accredited to the on die memory controller again.
Update: We copied the 32-bit marks from our benchmark in previous testing instead of the 64-bit. You can view the previous articles here from a month ago. The graphs have also been updated.
Facebook
Twitter
RSS
275 Comments
View All Comments
Locutus4657 - Wednesday, August 11, 2004 - link
#228: I don't beleive that for most people it's the inability to accept that A64 can be beat. The fact is there are certain tasks for which any P4 core is better suited and I think most people accept that. What I think most people don't accept are the faulty benchmarks and conclutions in the original artical. The one I personally questioned the most were the mysql benchmarks (corrected now). I'm sorry but the Prescott CPU is inferior in this type of environment as compared even to the old Northwood CPU! It has already been well established that the Opteron/A64 processors can best any Intel chip in DB envirnoments. So what this article was to have me belive is that a 12% increase in clock speed for Precott translates to a 60% lead in database benchmarks? Give me a break. Those test results alone were enough to call the entire original artical in question. I would like to add a disclaimer now that this issue and a couple others were fixed, so I would have to re-read the artical to make any sort of judgment about it again.eiger - Wednesday, August 11, 2004 - link
Snorre and Fritz64,
Thank you guys for your comments, It was
quite useful. We are probably going to go with
the Opterons as of now.
snorre - Wednesday, August 11, 2004 - link
I'm looking forward to the single CPU review with Xeon 3.6GHz vs. Opteron 150, and the dual CPU review with Opteron 250 ofcourse. I really hope these reviews will be much better, or else :P.porkster - Wednesday, August 11, 2004 - link
Fair review.I praise AMD for bringing 64bit cpu's to the home user but any business in their right mind would never purchase a clone/emulator cpu as of AMD.
Intel has some catching up to do with releasing 64bit to the desktop users at a good price.
I can't bring myself to purchasing an AMD cpu as I'm a programmer and also want future ready items like PCIe, BTX design, DDR2, etc which AMD based systems don't have.
Least there is sometime yet before 64bit OS and device drivers are available for the general market.
.
Viditor - Wednesday, August 11, 2004 - link
Aileur - "Is this review total crap because the opteron doesnt wipe the floor with the xeon?"I don't know if it's total crap because
1. it's in German and even Babelfish can't translate it very well
2. I can't find any listing of the setup used for testing. (given the proper setups, I can make even a K6 wipe the floor with Nocona...)
"As for the cpu difference question, thinking the 3500+ isnt targetted to be against a 3.5xx offering from intel is a nice try to convince everybody else, but im sure it doesnt even convince yourself"
1. it's not a 3.5xx but a 3.6xx
2. Targeting occurs between chips within a reasonably similar price range and market
fritz64 - Wednesday, August 11, 2004 - link
Hello eiger,For Scientific computing and best performance for your dollar, you will be in a better situation going with the Opteron. Are you looking for a dual processor set-up? If so, Opteron with on-DIe memory controller and NUMA (non Uniform memory access) aware architecture gives you the best performance again. Just like someone suggest earlier, go and read the bencmarks from aces hardware and get some feeling about the type of performance that you can get.
On the compiler issue, I will advise you to go for pathscale compiler suit for optimum performance on the opteron. if you choose to go for intel, the the Portlan group compiler or Intel compiler should be fine. A 30 day free trial version of the pathscale compiler can be obtained at www.pathscale.com.
You code seems to run faster (<10secons). I will advise you to task both system( Operon and Xeon) with a program that requires at leat >200MB of memory and ri=un for at least >30minutes. With that type of program all aspect of the hardware will be put to test. A code that runs for less that 10secons is probably running from the cache alone. Hope this information is useful.
allnighter - Wednesday, August 11, 2004 - link
reply to #220chainbolt wrote:
I guess you did not understand the intenton of this article in the first place?
______________________________________________
Oh God.
You know I'm trying to stay civil here so please do not insult my intelligence. Next time you comment on someone's comprehensive skills make sure your butt and your head switch places.
Aileur - Wednesday, August 11, 2004 - link
Baring in mind i am well aware this review is far from perfect, im wondering, how many reviews will it take for people to accept the opteron can be beat?Heres another one
http://www.tecchannel.de/hardware/1441/15.html
Granted, it is not a total wipe out, but it does win some and lose some. Is this review total crap because the opteron doesnt wipe the floor with the xeon?
As for the cpu difference question, thinking the 3500+ isnt targetted to be against a 3.5xx offering from intel is a nice try to convince everybody else, but im sure it doesnt even convince yourself. Whos gonna bench it against a tbird and find out "hey thats true! it is equivalent to a 3500mhz tbird!", ridiculous.
The xeon that was used in this review IS a prescott, the only differences are pins placement, and SMP support, so stop crying this review wasnt fair. Instead of wondering why they didnt use a "server" chip on the AMD part, realise that the xeon used IS a desktop chip, in a different suit.
And finally, for the name calling part, this site is free, they provide countless benches for countless pieces and theyre considered one of the most respected site on the net, so if youre not happy, you can go read one of those site that does 3 reviews a year, reviewing the stuff he just bought to put into his mothers pc.
-vent off
douglar - Wednesday, August 11, 2004 - link
The disappointing aspects of this review as I see them:1 - Questionable test setup. Where exactly was this Xeon64?
2 - Bad choice of benchmarks. He called them "synthetic benchmarks" but I don't even know if some of these tests should be called benchmarks at all when they only test putchar.
3 - Flawed testing methodologies. Let's see. Bungled compiles. Dissimilar test configurations. Copied results from other sources. Copying the wrong results from other sources. I don't think I've seen worse.
4 - Totally ignorant conclusions. The summary looked like a two bit fanboy site and was about as logical.
5 - Failure to address failures. Yes he did fix two of his more egregious mistakes, but Anand would have tacked on a full page disclaimer at the front of the article in red italics for something that was 1/10th as controversial. This guy just put a pair of "my bad" footnotes in the middle of the article and left the atrocious conclusion stand for us to enjoy.
cdo - Wednesday, August 11, 2004 - link
this so called review wont be an issue for much longer... we all know that aces and a myriad of other high tech sites will be doing this exact review over the course of the year. time will show just how bad a "review" this was.