Linux Database Server CPU Comparison
by Johan De Gelas on June 17, 2005 12:05 AM EST- Posted in
- IT Computing
Benchmarks MySQL 4.0.18: Intel versus AMD
It must be said that the MySQL results had a large margin of error (3% - 4%) compared to DB2, especially at high levels of concurrency.
Here is our MySQL configuration:
Read_buffer=2GB
Port=3306
socket = /var/lib/mysql/mysql.sock
skip-locking
set-variable = max_user_connections= 2000
set-variable = max_connections= 2000
key_buffer=2G
Read_buffer=2G
table_cache=1024
tmp_table=128M
max_heap_table=256M
read_rnd_buffer = 64M
thread_cache=16
net_buffer_length=16k
The " query cache" was off, as we wanted to test worst case performance. In some cases, the query cache was able to push a single Xeon to 1000 queries per second, and the CPU was still capable of doing more, as the CPU load was at 50% - 70%. At 1000 queries/s and more, other bottlenecks started to kick in, such as the latency of the network driver, the operating system and so on.
All numbers are expressed in queries per second. All concurrency tests below 5 are not reliable enough to make any firm conclusion as the margin of error is much higher.
| Concurrency | Dual Xeon (Gallatin) with L3 cache |
Single Xeon (Gallatin) with L3 cache |
Dual Xeon (Nocona) with HT |
Single Xeon (Nocona) with HT |
Dual Xeon (Irwindale) 3.6GHz with HT |
Dual Core Intel 3.2GHz |
Dual Opteron 250 2.4Ghz |
Single Opteron 250 2.4GHz |
Single Opteron 252 2.6 GHz |
| 1 | 243 | 248 | 280 | 277 | 286 | 233 | 290 | 298 | 319 |
| 2 | 357 | 317 | 423 | 338 | 450 | 344 | 438 | 370 | 399 |
| 5 | 466 | 356 | 473 | 358 | 497 | 442 | 543 | 435 | 470 |
| 10 | 505 | 361 | 521 | 375 | 517 | 487 | 629 | 465 | 502 |
| 20 | 496 | 350 | 531 | 371 | 545 | 507 | 670 | 455 | 498 |
| 35 | 508 | 355 | 555 | 371 | 506 | 490 | 665 | 470 | 507 |
| 50 | 497 | 348 | 526 | 368 | 495 | 502 | 669 | 472 | 508 |
| AVG | 494 | 354 | 521 | 368 | 512 | 486 | 635 | 460 | 497 |
| MAX | 508 | 361 | 555 | 375 | 545 | 507 | 670 | 472 | 508 |
Those were the raw numbers. Let us now analyse this...
| Concurrency | Dual versus Single Xeon Galatin | Dual versus Single Xeon Nocona/ Irwindale | Dual Opteron 250 vs Single |
| 1 | -2% | 1% | -3% |
| 2 | 12% | 25% | 18% |
| 5 | 31% | 32% | 25% |
| 10 | 40% | 39% | 35% |
| 20 | 42% | 43% | 47% |
| 35 | 43% | 50% | 41% |
| 50 | 43% | 43% | 42% |
| AVG | 40% | 41% | 38% |
MySQL ISAM is an incredibly fast database engine in our benchmark situation: it handles the same workload about twice as fast as DB2. I have to emphasize "our benchmark situation" because we cannot forget that our workload is mainly about reading the database and not writing. And of course, it must be said that the MySQL ISAM engine does less work on each query than DB2; it does not support transaction-safe (ACID compliant) commit, rollback, and crash recovery capabilities.
MySQL, as we have also noticed 6 months ago, doesn't seem to scale as well as DB2. At best, you get a 40% - 45% performance increase when the concurrency level is high enough. When we move to quad CPUs, we only get a 20% - 30% increase while DB2 still offers a 70% increase. The better scaling of DB2 means that with enough CPUs, it runs almost as fast as the MySQL ISAM engine, and offers all the transaction-safe capabilities as a bonus.
Let us check if the architectural differences between the CPUs make a difference . Again, don't pay too much attention to the results of the lower concurrency levels.
| Concurrency | Dual Xeon Irwindale versus Nocona (3,6 GHz) | Xeon Nocona (3,6 GHz) vs Galatin (3,06) | Opteron 2.6 vs Nocona 3.6 | Opteron 2.6 vs Pentium-D | Xeon Nocona 3,6 GHz vs Pentium-D |
| 1 | 2% | 12% | 15% | 37% | 19% |
| 2 | 6% | 7% | 18% | 16% | -2% |
| 5 | 5% | 1% | 31% | 6% | -19% |
| 10 | -1% | 4% | 34% | 3% | -23% |
| 20 | 3% | 6% | 34% | -2% | -27% |
| 35 | -9% | 5% | 37% | 4% | -24% |
| 50 | -6% | 6% | 38% | 1% | -27% |
| AVG | -2% | 4% | 35% | 2% | -24% |
| MAX | -2% | 4% | 36% | 0% | -26% |
The bigger L2-cache of the Xeon Irwindale did nothing more than compensate for the slightly higher latency of the L2-cache. The Xeon Irwindale and Nocona perform alike.
MySQL, unless you get the special Intel Compiler optimized version, remains the stronghold of the Opteron. The fastest (single core) Opteron outperforms the best Intel CPU by a 35% margin. We didn't use the Intel compiler version as we have reason to believe that this version is not used a lot in the real world. We might try it out in a future article.
The relatively limited scaling also means that high clocked single CPUs can be an interesting option. This is illustrated by the Opteron 252 2.6 GHz, which outperforms the dual core Pentium-D 3.2 GHz by a small margin.
Facebook
Twitter
RSS
45 Comments
View All Comments
Guspaz - Friday, June 17, 2005 - link
Why are there no graphs like other Anandtech articles? Why is everything in hard to read tables with broken formatting? This one seems a bit rough around the edges compared to the usual Anandtech quality.juhl - Friday, June 17, 2005 - link
I find it quite odd that you claim to be testing with a 2.6.12 Linux kernel despite the fact that that kernel has not yet been released in a final version.If you are using one of the pre-release kernels you should explicitly say so, and tell us which one.
The latest stable kernel at the time I write this is 2.6.11.12, the latest development kernels are 2.6.12-rc6, 2.6.12-rc6-git8 & 2.6.12-rc6-mm1 . There's also the question of wether or not you used a stock kernel.org kernel or a "patched to hell-and-back with crap" gentoo kernel...
So, what were you really using?
sinisterDei - Friday, June 17, 2005 - link
#20Correct me if I'm wrong, but would not the query cache positively affect the scores of both vendor's chips?
I suppose I don't have a pair of database machines just sitting around to test it out, but I'd imagine that if query cache was enabled the Opteron would experience similar performance boosts to the Xeon- if not more of a boost thanks to the higher-performing memory subsystem.
Just my $.02 to counteract the fear-monger :)
Viditor - Friday, June 17, 2005 - link
"Translation: We didn't want our beloved AMD to lose, so we doctored the test"Translation of the translation...
Shoot the messenger! :-)
michaelpatrick33 - Friday, June 17, 2005 - link
#20 Ah yes the conspiracy theories begin. Just like AMD with Tomshardware. The server results here appear pretty consistent with every other server test I have seen on review sites but who knows.# 19. Intel is only at 90nm but do have 300mm wafers. That is why Fab36 is so important for AMD. 300mm wafers and 65nm by Q2 of 2006 should put them pretty equal with Intel's fabrication level. Production level is still way, way in favor of Intel though.
Pricing, as I said before the Opteron dualcore chips are way cheaper than Intel dualcore server chips because Intel doesn't have any.
Questar - Friday, June 17, 2005 - link
"The " query cache" was off, as we wanted to test worst case performance. In some cases, the query cache was able to push a single Xeon to 1000 queries per second, and the CPU was still capable of doing more, as the CPU load was at 50% - 70%. "Translation: We didn't want our beloved AMD to lose, so we doctored the test.
thegagman - Friday, June 17, 2005 - link
BTW guys, one reason why AMD may be pricing its chips mihc higher is the MFG process. Unless I am mistaken (and someone correct me if I am wrong), they are still using 200mm wafers on a 90 or 110 process. Intel is using 300mm at 65 nm...this results in a huge difference in throuhput. Since AMD is already pricing its CPUs very agressively to gain market share, and the die of those dual-cores is much bigger (anybody know the real %?) then it is to be anticipated that their dual-cores are much more exspensive. They are probably gambling on selling dual-core Opterons at high-margins via Sun and other OEMs first,which will probably take most of their wafers. This is why their Desktop parts are coming later I would bet...thegagman
nserra - Friday, June 17, 2005 - link
#17 Good answers but, "Depends on the applications you run. On single threaded code, the faster single core will run your code faster."Doesn’t explain what I was meaning, so for you is OK to pay more for a single core processor, because runs faster some times (may run slower other times), how will you know what it will happen? (Supposing that you don’t know shit about your software requirements).
. AMD don’t have this problem so why would amd for example release an 4200+ processor at the same price of the 3500+ ? If the performance is equal or superior?
I think amd have made they right decisions, like intel have made his.
They all play with what they have, and not with they haven’t.
fitten - Friday, June 17, 2005 - link
"Explain me something:- how do you explain or how Intel will explain that their single core processor cost more than the dual core ones? "
Because Intel sets the prices of their chips. They want to push dual-core to the masses so they price them accordingly.
"- Why should you buy a single core over a dual core if it cost more? "
Depends on the applications you run. On single threaded code, the faster single core will run your code faster.
"- How good is this Intel market decision (marketing). "
Probably pretty good. Considering you have to buy a new motherboard to use the dual-core Intel parts, they dropped the price so that the CPU + motherboard cost is about the same (or less than still) the cost of just the dual-core CPU from AMD. Sounds like a good strategy to me.
michaelpatrick33 - Friday, June 17, 2005 - link
I also think it is safe to say that Amd's dualcore Opterons will be cheaper than any Intel dualcore server chip for the next six to eight months since there aren't any Intel dualcore server chips. IDC just released market research that showed AMD with 30% of the 4way server sales in Q1 '05. That is what AMD is after. The 64bit performance difference is surprising to say the least.