HyperTransport and Opteron/Athlon64 Overclocking
The first question many will have about our efforts to look at how Athlon64 will perform is how we can possibly compare an overclocked Opteron to a chip that is not overclocked. In the case of the Opteron, the comparison is more accurate than you might first think.In normal setups (e.g. Athlon/P4), the CPU gets its clock from the FSB clock and multiplies it by the “clock multiplier” to determine how fast its internal clock should be. With a 16x multiplier, when the external clock ticks once, the CPU ticks 16 times. However, with the Athlon 64/Opteron, there is no FSB, so the CPU must get its clock from somewhere. It doesn't produce it internally; instead, it derives it from the native HT (HyperTransport) frequency, which is 200MHz, but because of the bus' nature, it runs at an effective 800MHz.
So, for our 1.8GHz Opteron 144, the multiplier is 9x, which is why raising the HT frequency to 222MHz increases the clock speed to around 2GHz. But we are increasing the HyperTransport clock in our overclocking, and not a FSB clock, which does not exist on Opteron/Athlon64. In real terms, this means our CPU overclocking has a significant impact on Performance, but it is unlikely that our increase in memory speed will have nearly as much impact on performance. Since we are nowhere near saturating the Hypertransport bus at 200 (effective 800), increasing HyperTransport to 222 (888) will not likely have much, if any, impact on overall performance. Our performance improvements, with Opteron/Athlon64, are mainly coming from increase in CPU clock — much more so than on the Pentium 4 or Athlon architectures.
Obviously, the PCI bus operates at a different frequency than the HT bus than the CPU, but they all operate based on multiples of each other, and are all derived from the HyperTransport clock.
Facebook
Twitter
RSS
79 Comments
View All Comments
Anonymous User - Saturday, September 6, 2003 - link
Not to worry, there will soon be more competition for Opteron in the form of Low Voltage (and price) Itaniums, Prescotts, and even 800 MHz FSB Nocona XEONs. It's going to be very fun in the next year or so.Anonymous User - Saturday, September 6, 2003 - link
Am I the only one concerned that each test platform seems to use different amounts of system memory?Anonymous User - Saturday, September 6, 2003 - link
1 OF 2 things will have to happen. Either MS reduces their prices to compete with Linux or Linux will start charging/more for their OS. Open source is great and cheap right now which is why it is popular, but someone will try to commercialize it.Anonymous User - Saturday, September 6, 2003 - link
Fortunately in the next few years many folks will be switching to Linux both 32 and 64-bit just to get away from Windoze and all the stability and security issues with virtually every version of Windoze, bar none. This will be good news for AMD and the Opteron/A64 which both run very well on Linux and 64-bit Linux is available to all right now.Once the software companies wake up and smell the coffee and pull their heads out of Microsofts's butthole, they'll start releasing apps for Linux that look and feel like those for Windoze. This will facilitate a relatively painless transition for millions of folks who would switch to Linux immediately if they could import all of their existing apps files without headaches. Thankfully with enterprise and World governments switching to Linux, there is a clear financial incentive for software makers to get their act together and fill customer needs. The World will be a much better place when consumers have the ability to purchase a quality O/S and software apps and at this stage of the game Linux is a clear winner over any Windoze O/S for stability, performance and security.
Anonymous User - Friday, September 5, 2003 - link
Well when the P4 first came out it was slower than the P3, and costed around $1200-1500. You expect a new series processor to be dirt cheap? Yeah right.Prescott is what, a P4 with 20w more heat dissapation and 1MB cache and several (currently) useless instructions? OK, it has some other secret features. How long though can the 2x ALU stay 2x? They will run at 6.8ghz, and getting them to work at higher speeds will be 2x as hard as the rest of the processor (and it's now supposively going to be 4x?).
Of course, I guess everyone forgot, this is just a preview, not the acutal thing, so why get so worked up over it?.
Whether the desktop version is great or not, the more Opterons you run together only get better, with the Xeon you get less in return from going 2 to 4.
Anonymous User - Friday, September 5, 2003 - link
What is the reason that overclocked Opteron 244 is used instead of the real Opteron 246 that is available in retail? What is the reason to use 2 Gig of memory and compare it to a system that uses 512 MB memory? If memory doesn't matter in those bechmarks then why 2x256 isn't used?Why not to compare to Pentium 3.2 which is the top Intel desktop chip instead of 3.0 GHz?
Anonymous User - Friday, September 5, 2003 - link
Why on earth are the benchmark results in FLASH?Thats just really annoying.
Anonymous User - Friday, September 5, 2003 - link
This benchmark is pretty funny, it leaves a lot to be desired from todays reviewers. Not that I think the athlon64 isn't a very good improvement, but the results they are showing do not match any other results people have seen, reminds me of the Hardocp reviews.Anonymous User - Friday, September 5, 2003 - link
Just for your information number 49, the 2.0GHz CPU anandtech tested won't be the fastest CPU AMD releases on the 23rd, so comparing it to a 3.2GHz CPU would actually have been unfair.Anonymous User - Friday, September 5, 2003 - link
That's not anand hahahah :p