Socket 939 Roundup: Battle at the Top
by Wesley Fink on July 30, 2004 8:00 AM EST- Posted in
- Motherboards
Abit AV8: Overclocking and Stress Testing
FSB Overclocking Results
| Front Side Bus Overclocking Testbed | ||
| Processor: | Athlon 64 FX53 Socket 939 2.4GHz |
|
| CPU Voltage: | 1.5V (default) | |
| Cooling: | Thermaltake Silent Boost K8 | |
| Power Supply: | Antec TruePower 430W | |
| Maximum OC: (Standard Ratios) |
203 x13 2642MHz (+10.1%) |
|
| Maximum FSB: (Lower Ratio) |
280x9 (2520) at 1:1 Memory | |
Our FX53 easily achieved the next speed level of 2.6GHz on the Abit AV8. This will likely be the speed rating of the FX55, which will appear later this year. The best proof that the Abit has a working PCI/AGP lock can be seen in the fact that we were able to run DDR550 memory at 280x9 on the Abit. To achieve this overclock level, we needed to reduce the base HyperTransport frequency to 800. The AV8 could reach 300 CPU frequency at lower memory ratios and a 8 multiplier. We did find that the Abit did not really like an HT setting lower than 800, so we were limited to 300 as the maximum usable FSB.
Memory Stress Test Results:
The memory stress test measures the ability of the Abit AV8 to operate at its officially supported memory frequency (400MHz DDR), at the best performing memory timings that our Mushkin PC3500 Level 2 or OCZ PC3500 Platinum Ltd Modules will support. Memory stress testing was conducted by running RAM at 400MHz with 2 DIMM slots operating in Dual-Channel mode. The memory configuration of the Abit is a little unusual in that the first 2 DIMM slots represent a Dual-Channel. On most Dual-Channel motherboards, slots 1 and 3 are the first Dual Channel.| Stable DDR400 Timings - 2 DIMMs (2/4 DIMMs - 1 Dual-Channel Bank) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 10T* |
| Precharge Delay: | 2T |
| Command Rate: | 1T |
The Abit KV8 PRO was completely stable with 2 DIMMs in Dual-Channel at the best performing settings of 2-2-2-10, at 2.6V default voltage. It should be noted that the BH5 memory modules that we used for testing are no longer available for purchase, but we have not yet established our standard memory for future testing. We will be using a new standard memory in future motherboard tests.
As we first found in our latest Socket 754 roundup, the Command Rate is very important for top performance on a VIA chipset motherboard. The best performance is at a Command Rate of 1T, and the Abit AV8 was completely stable at a 1T setting with 2 DIMMs. Standard memory bandwidth measured with SiSoft Sandra 2004 SP2 shows a 6000 MB/s bandwidth with 1T Command Rate compared to a 5000 MB/s bandwidth with a 2T setting. While we could not set "Bank Interleave" in the BIOS, Sandra 2004 reported that a 2-way Bank Interleave was being used by the memory controller. With an on-CPU memory controller with Athlon 64, this is not as important a setting as it is on with chipset-based memory controllers.
Filling all four available memory slots is more strenuous on the memory subsystem than testing 2 DIMMs on a motherboard.
| Stable DDR400 Timings - 4 DIMMs (4/4 DIMMs - 2 Dual-Channel Banks) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 10T* |
| Precharge Delay: | 2T |
| Command Rate: | 2T |
The Abit AV8 was able to run with all 4 DIMM slots at the same aggressive 2-2-2-10 settings used for 2 DIMMs. However, Command Rate must be reduced to 2T when filling both Dual Channels.
During the course of testing for the 939 roundup, we experienced a failure of our FX53 processor. AMD was very gracious in quickly providing a replacement so testing could continue. The FX53 failure is probably rare, but the way that the CPU failed provided some insight into the Athlon 64 on-board memory controller. We first noticed new boards requiring much slower timings for memory; in most cases CAS 2.5 was required where CAS 2 had worked previously. This appeared to be motherboard differences, so we continued testing. Later, the FX53 failed. The replacement first went into a board where CAS 2.5 or 3 had been required, and now the same board performed fine at CAS 2 memory timings. Our point here is that the problem was the on-CPU memory controller failing and not the motherboard. The on-CPU memory controller is wonderful for reducing latency, but, unlike the past, memory problems may also be related to the CPU and not just the motherboard.
Facebook
Twitter
RSS
83 Comments
View All Comments
ThePlagiarmaster - Thursday, July 22, 2004 - link
#50 (RyanVM)...A quick check at pricewatch shows you CAN'T BUY IT (the 3.4EE S775, nor the 3.6E 775). Nobody even advertises the S775 for sale at top speeds. The 3.6EE has not even been released yet, and the 3.4EE can't be bought (775's I'm talking). Of course anand could get some engineering sample as before (Intel Confidential written on all pics I've seen ANYWHERE), but are we comparing something I probably won't be able to get for 2 months or more? Prescott 3.4 anyone? Took 3+ months for these to FINALLY show up as a product I can buy (and they were announced Feb2nd I think, reviewers got samples march 21st or so))...I'd call that a bit of an early review, just kills sales for the competitor as people wait for Intel (nvidia 6800 ultra anyone?). If you CAN buy it, show it to me. Otherwise for all intents and purposes the 3.4EE S775 chip doesn't exist. Is that a bit clearer for you? Did you bother to read where I said only 10 places are even selling the 3.4EE for S478? Paper launches and whats AVAILABLE are two completely different things. AMD can probably show some benchmarks of their dual core shortly, but who cares if I can't buy it until mid 2005 (or later)?The 6800 Ultra has been out for what like 2 months. I still can't buy that either (nor sell it, which is pissing my customers off..."why don't you have it, I've been waiting 2 months"...blah). The prescott 3.4E just finally started shipping last week for crying out loud (15th I think it was..strangely that press release is GONE everywhere..hmmm). I think reviewers should either completely disregard products that can't be bought or at least dedicate a good paragraph of two to the track record of the last release, and how you'll probably be waiting for months for the one 'we're reviewing today'. I'd say the same about ATI's, but at least one place is selling its best for $999 (ROFLMAO...IN STOCK though..heh).
Note that Dell's 3.6E isn't expected until December! http://www.theinquirer.net/?article=17347 (5 months away?) I can see telling me about something I might be able to get next week, or maybe 2 weeks, but we're talking 3-5 month paper launches here. Do my statements make more sense to you now? Maybe I'm just tired of customers chewing my butt as if I could do something about it.
FactorOfTwo - Thursday, July 22, 2004 - link
#34, the ECC support question is less obvious than you might think.The specs for socket 939 boards I have checked either say "non-ECC", or don't say either way, except I learned today that the MSI K8N Neo2 Platinum does say it supports ECC.
The general lack of support is surprising given the Athlon64's on-board memory controller. AMD's specs do say the 939 CPUs support ECC, but that does not obviously imply that all motherboards do.
Some official specs that explicitly say non-ECC:
ASUS AV8 Deluxe:
http://usa.asus.com/prog/spec.asp?m=A8V%20Deluxe&a...
ABIT AV8:
http://www2.abit.com.tw/page/en/motherboard/mother...
Wesley Fink - Thursday, July 22, 2004 - link
#49 & #50 - There is indeed a Socket 775 3.4EE, but as Anand showed in his 775 launch article the 3.6 is actually a faster CPU. you can check his results at http://www.anandtech.com/cpuchipsets/showdoc.aspx?...Our motherboard testing showed the same thing, which is why we did not even bother to test with the 3.4EE. It is true there a few benchmarks where you pick up a few points with the 3.4EE, but the difference is generally small and the 3.6 is faster overall. Of course the 3.4EE is in addition a lot more expensive than the 3.6. Given the 3.4EE cost and lack of standout performance, we felt the fastest 3.6 compared to the fastest FX53 was the fairest comparison at the top.
It would have possibly made sense to include Socket 754 data for the 3400+ for instance, but that is a different socket and could not be tested in these boards. Those benchmarks are available in the 2nd Generation 754 motherboard motherboard roundup at http://www.anandtech.com/mb/showdoc.aspx?i=2063. Since those benchmarks were done only a few weeks ago we didn't repeat them, but we could include the K8N Neo (754) running a 3400+ for referenc in future charts.
We never had an FX53 754 in the motherboard lab so that direct comparison can't be done here. However, Derek did exactly that comparison in his 939 launch review at http://www.anandtech.com/cpuchipsets/showdoc.aspx?...
RyanVM - Thursday, July 22, 2004 - link
#49: "There is NO P4EE for S775"There very much is a Socket775 P4EE at 3.4GHz. Did you read any reviews at all before spouting that off?
ThePlagiarmaster - Thursday, July 22, 2004 - link
#22 TrogdorJWFirst of all, ONLY 1 place shows the 560 P4 "IN STOCK" and only 2 others as PRE-ORDER. Is it even fair to say this CPU exists? The 3.6EE doesn't exist yet either in any socket. I would have liked to see more cpu's like you, but the 3800+ would be the only real need here, as it is the exact price range as the 560 (but a A643000+ and P4 3.0 would be cool for the poor). However, when looking at the situation as a whole who'd pass on 64bit, TOTALLY UNLOCKED chip, and domination for around $90-100? If you're spending $700 on a CPU wouldn't you spend $788 to dominate, and completely obliterate once 64bit gets going (did I mention the totally unlocked part?)? If that $100 really means that much the 3800+ w/64bit is much better (not to mention gaming domination now). On another note, you can't expect Anandtech to compare a CPU that won't even go into S775 yet. There is NO P4EE for S775. How could they possibly throw in one of those? Even for S478 there are only 10 places selling them (3.4EE, again no 3.6EE exists) let alone how many in stock. Test a dead end 3.4EE S478? Better include a caveat then that states it's about to be EOL.
Why test Xvid when its not a real standard? Divx is barely getting going in DVD players, getting Xvid on them too (right now) is a shot at a dart board. I'd prefer they drop these tests totally and show a DVD9-DVD5 rip as I think people are more interested in that. I don't know anyone encoding to divx when dvd burners are $60. Who actually watches movies on their PC (not many)? Who wants to fight to get X revision of Y codec to work on their setop DVD player? I think many more people are joining netflix (etc) and DVD9-5'ing it. Or just downloading what someone else already did (9-5) in the newsgroups. Downloading from the newsgroups (what divx was made for pretty much) is pointless when you have to try to get [x version, at y resolution and z codec and xyz audio bitrate (heh, ran out of letters)] to work on your dvd player. The only way to get them to work 100% is buying a player and doing your own rips that you prove WORK on your player (nobody trading on the web does vids to your dvd players specs mind you). But why would you go divx/xvid at all if you're doing this yourself (and not sharing them) with retail dvds? Wouldn't you dvd9-5 it instead and end all compatibility BS? If you're talking quality, divx/xvid etc is NOT DVD Quality. No matter how you encode it. Burn a DVD and get it over with :)
RE: 64bits breaking 4gb. That's not what its about for us Desktop users. It's about the extra registers etc. We are probably 3-5yrs from 6-8GB on desktops (as norm). On desktops you've seen anands (etc) tests with 2GB or under but getting the EXTRA registers benefits. In a server review I'd expect 8GB or more tested. In a desktop review 4GB will make no difference in anything they'd probably throw at this review. Also, if you can afford GB dimms, you'll buy a server BECAUSE you're trying to break the barrier for X app that you use and have 16 slots probably for smaller sized dimms (at 1/2 the cost). Desktops have no more than 4GB support in ANY board on either side. We are only after the OTHER benefits of 64bit. The extra registers can be used with just a GB of memory. I'm not against having 4GB in my machine, but I don't see how it would change any of these tests. Just 2GB in one test setup could show us if anything would happen, no need for 4gb just yet.
I'd also like to see more games thrown at these machines. Use fraps to toss in more game types that don't have benching built-in. Painkiller, Perimeter, Ground Control 2, City of Heroes,Battlecry3 etc. CPU/Motherboard tests should look more like the videocard tests. On top of that, aren't we all downloading something in say Xnews/Agent etc while playing some game? How does that affect FPS? Pick a game and test the same group like that. Saturate your cable connection by firing up an easynews download or something, and benchmark a game. I'd think this would show NForce3 and Intel's chipsets at a MAJOR advantage as the Ethernet is NOT PCI. I'd bet tons of people download while playing games (online or not). Life sucks on PCI in this case.
One more comment/question. Where is the driver that scored 80+ in DRV and DX and dominated the rest of the workstation tests, even aquamark etc in this review: http://www.anandtech.com/mb/showdoc.aspx?i=2036&am...
We're talking 20-40% improvements and that was on a 5950 vs. this reviews 6800ultra. Seems I'd really want that performance for FREE. Where's that driver? 88 to 57 in DRV? SHEESH. How'd they do that? Where can we get it? I expected workstation domination on nvidia boards here.
Anemone - Wednesday, July 21, 2004 - link
Silly me on the MSI lan - new board using the NF3 250 LAN, right design, I didn't read carefully.My bad
Anemone - Wednesday, July 21, 2004 - link
I like them being interested in overclocking and speed :)Curious of Asus will put out a NF3 ultra board. Curious why MSI didn't use the NF3 Lan - since its better and lower in cpu useage.
Great review!
llamas - Wednesday, July 21, 2004 - link
How long do we have to wait for micro-ATX motherboards in socket 939?JKing76 - Wednesday, July 21, 2004 - link
Thanks Wesley! Can't believe I missed that.Richdog - Wednesday, July 21, 2004 - link