A Look At AMDs Socket AM2 Platform

Page 4 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Stop! These bullshits!
First angry with THG because they have taken defective CPU from AMD or it was taken from AMD's dump by someone else and brought to THG.This CPU is fake enough and you are all eating your friends in forums for this issue.
SO, WHY DID AMD delay Rev F chips from Q1 to Q2
BECAUSE there was a fairly nasty bug in the DDR2 controller.And AMD delayed AMD CPU's from March/April to JUNE!
And Why THG released this issue?
Answer is they wanted to take you in interest to them as they usually did!
THG is my best hardware inspector website but they have making mistakes sometimes at launching new reviews about CPU or MOBO's.
Thanks everybody and Don't care about memory is DDR1 or DDR2 and FSB.Processor is also de-effective for bencmarking and ID (cpu-Z).All reasons don't make any sense.We all have to wait untill June is reached.
 
I guess you'd rather resort to personal insults rather than talk about the facts. That's fine. Carry on. Clowns like you are a dime a dozen. I've seen many tards like you come and go on these boards. You're no different.

-mpjesse
 
LOL. Whatever man.

The change from Socket 478 to LGA775 was basically for two things:

PCI-E and DDR2. The FSB only changed for P4 EE. As for the other crap you mentioned, it could have been done without a pin revision.

As far as the chipsets go- that's a totally different issue. I'm not talking about chipsets, I'm talking about sockets.

-mpjesse
 
I'll have to agree with eliottes on 1 thing. Everyone is getting far too worked up over the review. Personally, I feel that the review is quite valid for the individual processor they tested. The results are completely consistant with the known information about the early AM2 processors that we know so far.

DDR2 800 is supported by the motherboard since that's relatively easy to program in and reinforce the connections, but it isn't supported by the OMC because the original AM2s were only designed to handle up to DDR2 667. The lack of support in the review is not a conspiracy.

Neither is the poor timings of the DDR2 667 or it's poor memory performance. Anand has long reported that their early AM2 sample suffered poor memory performance which was why they didn't post performance numbers with their platform preview. This was further corroborated by The Inquirer which reported that AMD has discovered a memory controller bug in the OMC crippling performance and this was partially responsible for the AM2 delay to June. The THG numbers are not out of wack from all the other information out about early AM2 performance.

The thing I'm stressing is that these are probably correct results for early AM2 results. How reflective on the final product will they be? Not significantly. Tom's sample obviously had the OMC bug, which AMD has already corrected and they've also taken the June delay to add DDR2 800 support. Instead of looking at the negatives of the review people aught to look at the positives. If AM2 can perform generally only slightly below S939 using 4-4-4 DDR2 667 and with a memory bug, just imagine what the difference DDR2 800 and the fixed OMC will make?

That said, I'm still not expecting vast performance increases even with DDR2 800. Double the bandwidth is nice, but that assumes that the processor will actually need it. Otherwise it'll just go to waste. I'd be amazed if AM2 will even use all the extra bandwidth of DDR2 667 because that would have meant that S939 was extremely bandwidth starved which I've never heard associated with the platform. The fact is, even with more bandwidth the core architecture is still the same and it seemed to be working quite well already on S939. AS MadModMike already said, the bandwidth of DDR2 will only be beneficial in server and maybe dual core processors. Single-cores won't likely see much benefit.

There's just too much criticism of the quality of the THG review when they've already taken great pains to not the pre-production issues with the article such as right before all the benchmarks on page 9, in the Test Setup, and in the Analysis. In terms, of the timing of the article, there really isn't any negative conspiracy behind it. All launches are always surrounded by NDAs, the fact that this article is released this far before the launch means that AMD must have approved it. Otherwise, THG would already be under a lawsuit. Besides, I doubt that only THG and AnandTech has samples. If THG didn't review it someone else would have. It would have probably save the authors quite a bit of grief too.
 
On the issue of the CPU-Z screenshot:

I briefly exchanged messages with Patrick Schmid (the author). He believes that CPU-Z is incorrect in stating the memory is running at 100.5Mhz and will update the review to include a caption of the screenshot.

Clearly though, if the RAM was truly running at 100mhz the difference in performance would be very great.

So it appears (to me at least) that CPU-Z is incorrect.

-mpjesse
 
Bunch of bickering school children.

This is an engineering sample of a proc not to be released on the market for some months yet.

Lets look at this preview and see what positives we can find:
- AMD have finally moved to DDR2.
- Despite 2/3 posts being undecided on the memory timings, etc overall its on par with current AMD platform performance (i.e. no real performance drop in their initial move to DDR2).
- its stable (I think everybody missed that bit)
- Its only going to get better.

That's all I'm taking from this article. AMD have moved on a step and its all looking good so far.

Me personally, I'm happy to wait for a DDR2 65nm AMD proc before I upgrade.
 
I'm talking about sockets.
Click to expand...

The 775 shift not only included DDR2 and PCI-E, but there is a significant performance difference between Intel pentium 4 3.0 and Pentium 630. Have a look at Tom's CPU Charts for multitasking for instance.

,,
 
Bah I'm just happy that AM2 hit S939 with this memory when you can go grab DDR2-800 1GB sticks running at 4-4-4-12 for $125... just wait until manafacturers actually start pumping out DDR-800 for the mainstream instead of the select few Intel Entusiasts.

Anyways what about socket F thats whats really interesting, when are we going to get some good info on it?

*EDIT* Oh yeah Nvidia forum age doesn't matter take it from a guy 3 days older than 3
 
This wont be AMDs huge movement. They are most likely holding their trump card for 65nm. The main movement in this is to sopport DDR2 wich will allow more custimization. Also the power usage is way down which is good news for desktops but even better for there mobile products. If there desktop CPU can get 35 watt, what will there mobile cores be?
Although i was hoping for a performance boost, I cant help but think AMD is holding back for 65nm. I guess we will have to wait and see.
 
most mainstream customers want quiet little beige boxs the can stick under their deska, and amd is moving twords lower wattage as to increase the processors in the mianstream market which is dominated by intel mostly in dells.

but doesent a 5000+ just sound cool?

cant wait till it comes out so i can get good s939 stuff for dirt cheap!!!
 
im actually impressed - performance is about the same, yet the thermal specs are really impresive (35w for the x2 3800+ etc) - i wasnt expecting a performance boost so much.

When conroe comes out it might still be hotter but if its "20% faster" as they claim it will be valid for the heat, compared to the prescotts which now took like a total joke - a 3800+ X2 - 35w is not even 1/3 the heat of a prescott or smithfield and the prescott doesnt even have an IMC (comparing 90nm chips).
Click to expand...

the new revision preshotts (90nm) only put out 84.0w btw..
 
Yes and you're forgetting the different sizes in cache. P4 630 has a 2MB L2 cache (LGA775) while the 3.0Ghz Prescott (socket 478) only had a 1MB L2 cache. Once again, we're speaking about performance differences in the processors- not the sockets.

🙂
 
low latency is a must for K8 performance,

I don't think the result is fake, based on such a high latency, K8 CPU can't do anything

AMD sure will lose this time if the high latency problem can't be solved.

better save some money for Conroe.

P.S. I am a Opty165 user (o/c 2.8GHz).
 
I am sorry Patrick, but I think you have been had.
It's not like your employer is a close friend of Hector and the boys.
Dont you wonder why you got this sample?
Do you really believe that Sandra would score that low, if you were actually using a 333mhz memory bus?
For that matter, what timings were you getting? Was it 1T, 2T or maybe even 3T?
You have been used. Amd wants to give Intel into a false sense of security.
You should be offended.
 
low latency is a must for K8 performance,

I don't think the result is fake, based on such a high latency, K8 CPU can't do anything

AMD sure will lose this time if the high latency problem can't be solved.

better save some money for Conroe.

P.S. I am a Opty165 user (o/c 2.8GHz).
Click to expand...
Then explain to me how a Pentium 4 with DDR2-533 can score better in Sandra than this Athlon64 supposedly running DDR2-667. Or are you saying that Intel's off-chip memory controller is somehow more "suited" to DDR2?

So far, I'm seeing so much BS in this thread that it isn't even funny.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom