I was looking in the bios for the specs on my A64 3200+, and the FSB is only listed as 200 mhz. I'm still a n00b to a lot of this tech stuff, but I was under the impression that the A64s had a FSB of like 1600 mhz or something like that. Do I need an update to my bios, or is this normal? I'm really paranoid about my computer parts, because in the past I have had strange performance problems with various systems, so I am trying to make sure this machine is running correctly.
Athlon 64 FSB
- Thread starter Atomsmasher544
- Start date
200 mhz multiplied by the hypertransport speed of 800 mhz = the marketing figure of 1600 mhz. In reality, its a 200 mhz fsb, just like the 800 mhz fsb on a P4 is also just a "quad pumped" (marketing ploy) 200 mhz fsb.
200Mhz is the speed of the RAM, using DDR it is doubled to 400MHz. The speed of the Hypertransport link on the CPU is 800Mhz but as it is full duplex (send data both ways) it is sometimes quoted as 1600MHz
I also am wondering about ram speeds, I took a screenshot of my comp info with CPUz
http://www.geocities.com/atomsmasher544/pics/cpuinfo2.jpg
My ram numbers seem different than what other people list, is it running slow?
http://www.geocities.com/atomsmasher544/pics/cpuinfo2.jpg
My ram numbers seem different than what other people list, is it running slow?
Bingo. Here is a little lesson for yall. FSB is the pipe between the cpu and the memory controller. Because the athlon 64 has an on die memory controller it has no fsb. This means that no matter how fast intel makes their fsb, it will always be slower then hypertransport. The 200mhz your seeing is how fast the cpu communicates with the memory.
Sorry to burst your bubble but 800x200 doesn't equal 1600. It = 160,000 which is just a random number. The 1600 "FSB" is 800 MHz hypertransport "dualpumped" by running in full duplex to give 1600. It doesn't affect your BIOS settings. You're seeing the RAM speed of 200 MHz.
OK, sort this out once and for all!
1) The 200MHz is a bace frequency and everything else is a multyple of that.
2) There is no FSB, the memory controler is on die.
3) Hyper transport opperates at 200 MHz but is 'quad pumped' to give 800MHz, since data can flow both ways at 800MHz effective frequency it is somtimes refured to as 1600MHz = 2 * 800MHz.
4) As reguards the CPUz screenys, everything is in order, tho your latancys of 3,3,3 will be slowing things down a lot, if you had some better ramand ran it at say 2,3,2 for eg, you would see a noticable improvment. The memory is opperating at 400MHz (200*2 for DDR) which is correct for PC-3200 RAM at defult settings.
5) If you want to try some overclocking you can gradualy rase the 200 MHz base frequency and all the other frequencys (HTT, RAM, CPU and unfortunatly (since no lock oon current mobos) PCI and AGP) will rise by a factor deterined by there multyplyer.
Hope that helps clear things up, basicly your system is properly configured and there is no need to worrie unless you are getting crappy performance.
Good luck.
1) The 200MHz is a bace frequency and everything else is a multyple of that.
2) There is no FSB, the memory controler is on die.
3) Hyper transport opperates at 200 MHz but is 'quad pumped' to give 800MHz, since data can flow both ways at 800MHz effective frequency it is somtimes refured to as 1600MHz = 2 * 800MHz.
4) As reguards the CPUz screenys, everything is in order, tho your latancys of 3,3,3 will be slowing things down a lot, if you had some better ramand ran it at say 2,3,2 for eg, you would see a noticable improvment. The memory is opperating at 400MHz (200*2 for DDR) which is correct for PC-3200 RAM at defult settings.
5) If you want to try some overclocking you can gradualy rase the 200 MHz base frequency and all the other frequencys (HTT, RAM, CPU and unfortunatly (since no lock oon current mobos) PCI and AGP) will rise by a factor deterined by there multyplyer.
Hope that helps clear things up, basicly your system is properly configured and there is no need to worrie unless you are getting crappy performance.
Good luck.
humm... I saw a fairly noticable improvment moving from 3-3-3 to 2-2-2 with my system. I got about a 10% inmprovment in UT2003 and also the 'feel' of the system was mych improved, the general responsiveness seemed much better (seeing less of the windows hourglass and that quick 'flicker' when alt-tabbing out between apps dissapered). I know that that is a bit subjective, but i believe it made quite a diffrence and am happy with my low latancy ram purchase.
I not sure ware your 2% figure is from, but if it from a P4 system, the A64 benifits much more from low latancys. Also, the benifit changed depending on the application/benchmark being run.
Anyway, I just thought that may be what Atomsmasher544 menat by "My ram numbers seem different than what other people list, is it running slow?" Still, biostud is right, the performance diffrence is not that huge.
I not sure ware your 2% figure is from, but if it from a P4 system, the A64 benifits much more from low latancys. Also, the benifit changed depending on the application/benchmark being run.
Anyway, I just thought that may be what Atomsmasher544 menat by "My ram numbers seem different than what other people list, is it running slow?" Still, biostud is right, the performance diffrence is not that huge.
Agree 100%. I don't know where is that misconception about low latencies not being important coming from. It is very, very important. It is easy to see that on benchmarks, I have run Athlon64 in all (im)possible combinations with memory and my first hand acount is that you would see a noticable improvement between 3-3-3 and 2-2-2, sometimes up to 15-20%, give or take. It is extremely important if you are able to have aggressive timings.
Jeff7181
Lifer
- Aug 21, 2002
- 18,368
- 11
- 81
I was under the impression that because tha A64's memory controller is build into the CPU that the timings of the RAM didn't play THAT big of a role... got benchmarks to prove otherwise? If not, could you do some to prove this? Though I can't find them, I thought I remembered seeing benchmarks that show there's hardly any difference between 2-2-2-5 and 3-4-4-8
You're kidding? I am a lazy man to post pictures and stuff but I have about 15% improvement on my Fritzmark (chess) when 3-3-3-4 is compared to 2-2-2-5. I was using Micron chip at first and then I replaced it with BH-5. Huge difference. Extremely important to me when playing chess.
Whoever here has Athlon 64 just try SuperPI on 4 MB or 2 MB and see the difference between 3-3-3 and 2-2-2.
There are some nbenchmarks you won't see the difference, say Sandra CPU benchmarks come to mind but do not be deceived, in the real world applications (in my case chess) the difference is serious enough.
The benchmarks ware you see the biggist diffrences are those with highly branched code... AI in games for example. A lot of artificial benchmarks will not show this and some will exagerate it, my point was that the diffrence is noticable.
- Aug 22, 2001
- 31,886
- 32,092
- 146
Timings for gaming on A64 aren't very important and the gains in scientific apps is better but nowhere near 20%. The best reason to get tight timing ram is that by relaxing the timings you can hit higher overclocks on average.
Tom's
Ace's
PCStats
techwarelabs
Anandtech article "Third, the development of DDR2 memory, which at DDR533 is being launched with 4-4-4 to 5-5-5 timings. An interesting development with DDR2 is that while memory timings are even slower than the worst DDR timings, the performance can sometimes keep up with DDR because of other internal memory timings. This is not praise for DDR2, but all these developments bring to the forefront that CAS Latency is not the only thing that matters in memory performance, and in fact, it may not even be the most important specification as we move to faster memory. This shows the playing field is changing so you old dogs will need to learn new tricks so to speak. The mantra of 2-2-2-5 being a big boost over 2.5-3-3-6 is bullsh*t
Tom's
Ace's
PCStats
techwarelabs
Anandtech article "Third, the development of DDR2 memory, which at DDR533 is being launched with 4-4-4 to 5-5-5 timings. An interesting development with DDR2 is that while memory timings are even slower than the worst DDR timings, the performance can sometimes keep up with DDR because of other internal memory timings. This is not praise for DDR2, but all these developments bring to the forefront that CAS Latency is not the only thing that matters in memory performance, and in fact, it may not even be the most important specification as we move to faster memory. This shows the playing field is changing so you old dogs will need to learn new tricks so to speak. The mantra of 2-2-2-5 being a big boost over 2.5-3-3-6 is bullsh*t
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 21K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter