Linux and the Desktop Pentium M: Uncommon Performance
by Kristopher Kubicki on December 24, 2004 12:00 PM EST- Posted in
- Linux
Summing it all up
Overall, Dothan provided us with some sporadic, but interesting, performance gains and losses. Unfortunately, Pentium M just doesn't scale similarly to Pentium 4 or Athlon 64 in any application, although it does seem to mimic the performance of one or the other occasionally. On our OpenSSL tests, Dothan continually out-nudged even our mid-range Athlons, but then fell far behind in compilation and some content creation tests.There are, however, bottlenecks in the performance. High speed memory is something that our Dothan severely lacked on Linux, and we would certainly appreciate the next generation Alviso chipset to support something a little faster than DDR400. However, as Pentium M is a notebook chipset first and a blade/desktop chip second, the demands of low power notebook memory certainly take priority over a niche SFF/HTPC crowd.
The first surprise in our analysis came with the SQL database tests. Our windows benchmarks have shown in the past that the additional L3 cache can be quite helpful for database applications, and the 2MB L2 cache found on the Dothan plays a huge part in boosting performance. On the other hand, the additional cache might have been the same reason why GCC performed so poorly - although we hope that the Linux compile test was just a fluke (Update: Please see the note on the Compiling page. We believe we had an isolated fluke with the PATA driver that limited our performance). Other benchmarks put Dothan right in the upper middle of the pack, usually beating out the Pentium 4 offerings, but occasionally beating out the best that our Athlon 64s could produce as well.
Dothan isn't the miracle chip that we would have liked it to be. For starters, it is horribly expensive still. The 2.1GHz Dothan that we previewed today runs at around $500, and the motherboard costs another $270. For just a barebones configuration, our Pentium M desktop runs at around $1000. Granted, the overclockability on Pentium M seems outstanding, but finding slower, cheaper Dothans in socket 479 pin configurations may be a problem.
Unfortunately, we are only getting a small glimpse of the story here today. Our preliminary benchmarks on Windows show that Dothan does some awesome things on Windows; the compilers and operating system get a little more help from Intel in the design phase. Unfortunately, the extremely powerful and free Linux compiler remains dully unaware of many of the benefits that Pentium M has to offer, and as a result, it gets hurt painfully under the default or wrong compile flags.
All in all, Dothan does some very exciting things. The promise of cool, efficient powerhouses - from Intel, nonetheless - certainly has our attention. We will be keeping a very close eye on Pentium M over the next few months, particularly with the upcoming Alviso launch. If Dothan's Linux performance keeps up this well on the 855 chipset, we can't wait to see what it does on faster memory and the 915 Northbridge.
Facebook
Twitter
RSS
47 Comments
View All Comments
- Saturday, October 24, 2009 - link
http://www.goph3r.com/mh">http://www.goph3r.com/mh(air jordan, air max, shox tn, rift, puma, dunk sb, adidas) nike jordan shoes 1-24 $32
lv, coach, chane bag $35
COOGI(jeans, tshirts, hoody, jacket) $30
christian audigier(jeans, tshirts, hoody) $13
edhardy(shoes, tshirts, jeans, caps, watche, handbag) $25
Armani(jeans, tshirts,) $24
AF(jeans, coat, hoody, sweater, tshirts)Abercrombie & Fitch $31
http://www.goph3r.com/mh">http://www.goph3r.com/mh
BullCreek - Thursday, January 13, 2005 - link
Does anyone know where I can buy socket479 dothan core pentium Ms online? Newegg and various other places stock the socket 478 variety, but I haven't found anyone that has the socket 479 type.vl - Thursday, December 30, 2004 - link
If you are using an IDE drive make sure DMA was enabled. To check you can use the hdparm command.Lynx516 - Tuesday, December 28, 2004 - link
Something is disasterously wrong with your setup. Have you tried using a live CD and try compiling with that?What command are you using to compile the kernel? Are you sure you have GCC installed correctly? What do you get if you type gcc-config -l?
Are you sure nothing else is running (well major tasks)?
Have you tried running it as root?
KristopherKubicki - Tuesday, December 28, 2004 - link
Correct they are "user". Here are the results with GCC 3.4.3 on the same system on the second run.real 20m42.192s
user 16m39.148s
sys 1m21.287s
dave:~/bench/gcc/linux-2.6.4 # cat /proc/cpuinfo | grep MHz
cpu MHz : 2105.068
... Not sure what is the problem here. The CPU clock is correct. If anyone else has any ideas as to why only the GCC benchmark would be affected I am all ears.
Kristopher
Yozza - Tuesday, December 28, 2004 - link
Results with gcc version 3.4.3 20041125 on the Banias 1.7GHz (DDR266/400FSB).run 1:
real 4m40.758s
user 3m58.039s
sys 0m20.060s
run 2: (after "make clean")
real 4m24.759s
user 3m58.409s
sys 0m19.978s
run 3: (again, after "make clean")
real 4m27.728s
user 3m58.233s
sys 0m20.161s
Of course, this is with gcc 3.4.3. I'd rather not install 3.4.1 on this system only to rollback to 3.4.3 straight after.
I assume the kernel compile times you showed in the article were from "user" (or user + sys), rather than "real", as the latter is affected by other things happening on the system whereas the former gives the actual CPU time taken by the process.
The results are impressive to say the least. It's certainly extremely puzzling that your results with Dothan were so disparate. One would expect it to do very well indeed based on my results with 1.7GHz Banias + DDR266.
For comparison, my 3GHz NW P4:
run 1:
real 4m50.687s
user 4m19.365s
sys 0m25.261s
run 2:
real 4m46.675s
user 4m23.775s
sys 0m24.567s
KristopherKubicki - Tuesday, December 28, 2004 - link
Yozzo, Can you please do kernel 2.6.4 like we did in the benchmark?Kristopher
Yozza - Tuesday, December 28, 2004 - link
Development on the 3.4 gcc branch has only recently stabilised, and the pentium-m cflag was indeed quite buggy/broken in earlier releases, hence why I stated that gcc 3.4.1 is "pretty old". Not in the absolute sense, but rather in terms of the amount of development work that has gone into the 3.4 branch since.Some quick benchmarks with gcc version 3.3.4 20040623:
System was P-M Banias 1.7GHz, 768MB DDR266, i855PM; on a Dell D600.
# time make
...
real 9m36.940s
user 8m40.395s
sys 0m39.643s
2nd run:
real 9m35.056s
user 8m40.515s
sys 0m39.664s
For comparison, my 3GHz NW P4, 1.5GB DDR400, i875P:
1st run:
real 10m0.508s
user 9m0.929s
sys 0m46.210s
2nd run:
real 9m58.564s
user 9m3.107s
sys 0m45.658s
Kernel was 2.6.10-ck1-nitro1, with custom .config (http://homepage.ntlworld.com/yaoyaoliu/.config).
I didn't have gcc 3.4.3 available on my laptop to test, but gcc 3.3.4 clearly shows the expected performance and hence that there is nothing inherently wrong with the P-M architecture that makes it slow at compiling -- indeed, the 1.7GHz Banias P-M (DDR266, 400FSB) is slightly faster than my 3GHz Northwood P4 (dual DDR400, 800FSB).
I was also interested to see how much faster gcc 3.4.3 was compared to 3.3.4, and so decided to do a compile on my 3GHz P4 desktop machine (same kernel):
real 6m34.443s
user 5m52.910s
sys 0m33.019s
Interesting results indeed.
KristopherKubicki - Monday, December 27, 2004 - link
The GCC we used is the same GCC bundled with SUSE 9.1, our test platform. I don't think 6 months is too old for a processor architecture that has been available for close to 2 years. However, we will get a chance to revist these numbers in the very near future, and i will redo the tests with GCC 3.4.3 and GCC 3.4.1Hope that helps,
Kristopher
vaystrem - Monday, December 27, 2004 - link
Fromhttp://gcc.gnu.org
The GCC Homepage:
3.4.1 released July 2004
3.4.3 released November 2004
If you look at the changelist there have been a lot of improvements even between these few releases.