A Look At AMDs Socket AM2 Platform

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
This article clearly states that these are NOT the results we are likely to see at the June launch, or even with Revision F next month.
Click to expand...
Which begs my question, what is the point of the article? A 'sneak peak' at a screwed up PC that in no way reflects real world results? :roll: That's not exactly useful or informative. :?
 
ok, a few things,
1.)if the new socket is pin compatable with 940/939 but slightly different, what is stoping a 3rd party from making an upgrade path to AM2
Click to expand...

For a second I almost thought this wasn't a half-bad idea. Seeing how they use the same chipsets, it's essentially the same processor but with a different memory controller and pin layout. So, if you created an AM2 to 939 device, that can override the AM2 CPU's memory controller with it's own DDR(1) controller, it could theoretically work. Now, the reason why it's impractical. If you take an AM2 CPU and use such a device, you essentially have the same exact thing as that CPU made for 939. The cost of the AM2 CPU + conversion device would be greater than the 939. Even if you considered it viable for an FX-62 or 5000+ X2, the people buying those can afford a new motherboard.

My mobo has a slot designed to upgrade to an AM2 compatable riser board...I had to email them to confirm this. Sorry there is no official link but hey..paid $46 for my mobo and it's the bestest :):):)
 
What I found to be ridiculous is their SiSoftware Sandra Memory bandwidth results, they put DDR2 667 at 5.8GB/s, but their past Benchmarks put DDR2 667 in an Intel platform at 6.6GB/s, hmm... 800MB/s loss, that aint due to different RAM, unless this RAM was special RAM...

~~Mad Mod Mike, pimpin' the world 1 rig at a time
Click to expand...

the timings, dude, the timings... I've seen tighter timings on Intel rigs... and I've tried relaxing the timings on my RAM, makes a difference in the bandwidth, and should even more so with DDR2 and dual channel
 
The only thing I can as to the point is to whet the proverbial whistle of the enthusiast crowd until something more concrete is known.

That, and the fact that most sites have been treating anything about AM2 as huge news owing to AMD being so close mouthed lately compared to Intel.
 
You can see from my words I am talking about probabilities based on actual data we have from AMD. Don't you think if AMD had better chips would send them to Tom’s? Most probably Intel’s statements are based on these prototypes. Finally I'd be happy to see better chips coming out of AMD and have nothing against that. The final winner is the consumer.

,,
 
The only thing I can as to the point is to whet the proverbial whistle of the enthusiast crowd until something more concrete is known.
Click to expand...
By annoying the enthusiasts with such a bad article that they can tear to pieces? :roll:

Or am I missing something :?:
 
Oh the humanity, oh the bloodshed. ITs all over engineering samples too. I think we can all agree this is not somethign AMD intends to release. They will fix it, and if they don't thier reputation they have worked so hard for will be shot to hell. Lets take it as it comes, its a preview of something that isnt due for 4 months. Thats alot of time to fix it, especially in the tech industry.

I for one am in favor of the switch to AM2, all be it slightly over due maybe, but thats a matter of opinion. It looks like a good setup for Vista release :)
 
I refuse to belive that a chief engineer working for Intel is stupid enought to estimate the performance of an unreleased future intel chip based on an equally unreleased and obviusly buggy early sample of an AMD chip. 8O
 
Uhm...if you looked at the test setup, you would know that they DID use the fastest RAM available:

"AMD Athlon 64 X2 4800+ Windsor (2400 MHz 64+64/512 kB)
Support for DDR2-667 ONLY! Production processors will support faster DDR2-800 RAM."

"Corsair CAM2X512A-5400UL (XMS5400 V1.2)
2x 512 MB DDR2-667 (333 MHz, CL3-2-2-8, 1T)
Running at CL-4-4-4-12 Timings - not changeable in test motherboard BIOS"

http://www.tomshardware.com/2006/02/21/a_look_at_amds_socket_am2_platform/page8.html
 
the timings, dude, the timings... I've seen tighter timings on Intel rigs... and I've tried relaxing the timings on my RAM, makes a difference in the bandwidth, and should even more so with DDR2 and dual channel
Click to expand...
Ohh, for Christ's sakes... You don't lose nearly half the bandwidth just because you're using 4-4-4 timings (which is perfectly normal for DDR2-667, by the way). Get real. Now, please.

Don't you think if AMD had better chips would send them to Tom’s?
Click to expand...
It wasn't AMD who sent Tom's Hardware this AM2 CPU.
 
The problem is that the "Early Engineer Sample" made by AMD is treating the fastest RAM aviable as a PC2-3200 at CL-4-4-4-12. There for this article states nothing about the expected performance of the future AM2 chip. As i said a few post earlier the only conclusion that can be drawn from this article is:

A buggy early engineer sample made by AMD 3-4 month ago is 1-2% slower than a fully tested mobo/processor combo at the same freqvency . Also they are using the same bandwith RAM sticks with almost double the cas latency.
 
You can see from my words I am talking about probabilities based on actual data we have from AMD. Don't you think if AMD had better chips would send them to Tom’s? Most probably Intel’s statements are based on these prototypes. Finally I'd be happy to see better chips coming out of AMD and have nothing against that. The final winner is the consumer.

,,
Click to expand...

Do you honestly believe everything Intel or AMD tells you about how well their future chips will perform compared to each other? If so, I feel bad for you, as you have probably been greatly diasappointed in the past.

A far as AMD sending a better chip to Tom's, did you read the part of the article where it stated that a newer pre-release version that is expected to fix some of these issues is going out to manufacturer's this week? Apparently not.

We should all recognize that this is a good preview of the tech, but it is BY NO MEANS CONCLUSIVE. A lot can change between now and June 6th, and its very likely that we will see very different benchmarks in the future.
 
The problem is that the "Early Engineer Sample" made by AMD was treating the fastest RAM aviable as a PC2-3200 at CL-4-4-4-12. There for this article states nothing about the espected performance of the future AM2 chip. As i said a few post earlier the only conclusion that can be drawn from this article is:

A buggy early engineer sample made by AMD 3-4 month ago is 1-2% slower than a fully tested mobo/processor combo at the same freqvency . Also they using the same bandwith RAM sticks with almost double the cas latency.
Click to expand...

Thank You. That's what I have been trying to say. This whole article is starting to seem more and more like fanboy bait.
 
All I been trying to say since the beginning is that those are not credible benchmarks, but everybody is saying "oh they're so true, just read the article" I've read that article once, all it took for my to determine it was BS. Windshear: higher latencies DO NOT affect bandwidth that much, either on an A64 or P4, the most you lose from 2-2-2-5 tight timings to 5-5-5-15 (the worst) is about 200-350MB/s, but you gain so much back going from standard 200MHz to 333MHz which is DDR2 667, there is no way in hell that those benchmarks are true, as stated before, so let this thread die now, it's proven that article was useless.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
 
Obviously from the competition, if Intel next generation improves by 20% in performance as they promised and AMD sticks to the current results as they have shown, it is not difficult to foresee the future.

,,
Click to expand...

Reminds me somewhat of AMD's situation when they were in the limbo between AthlonXP and Athlon 64. Having reached high enough clock speeds the P4 was showing significant performance advantages over AXP. Then AMD released the K8 and took back the performance crown. Now they've been riding that success for a while and it's really time for intel to turn the tables again. The ball will be back in AMD's court soon enough though, everything depends on what they do with it.
 
Ok I have a couple of questions, one a bit dumb and one less so.

1) These latencies are in clock cycles right? Is that CPU clock cycles or ram clock cycles? If it's ram clock cycles then surly a latency of 4 @ 800Mhz is equivalent to a latency of 2 @ 400mhz, is this right or am I missing something?

2) Does anyone know if AM2 will sort out this thing where you have to drop to a 2T command rate when running 4 DIMMs?

P.S. Don't take the piss too much, I know I’m dumb.
 
Ram Clock Cycles I believe, but the speed of the RAM speed is dependent upon the CPU frequency divided by the divider. As for being dumb, being dumb and not knowing the answer are 2 very diffrent things. Just ask if someone can explain it, don't short change yourself :)
 
Correction, the CPU Freq and RAM Freq have no bearing on each other, thus why you can have 2:1 and 1:2 RAM dividers. The latencies are in clock cycles, such as CAS, Column Address Strobe, the CAS latency is the amount of time it takes to retreive data on the Column of the grid of capacitors physically inside the chip. This is because the order in which data is accessed, the Column (CAS) is accessed last, thus this is the time it takes for the data to actually be retrieved, in a sense. What I mean by "retrieve", is how long it takes for the request from the Memory Controller to get to the data, and output it to the pins.

For some info: RAM is accessed first by choosing a bank, then RAS, then CAS.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
 
correct me if I am wrong but the upgrade to M2 is about being able to adress more RAM because with upcoming 64bit OS's and Applications more ram is going to be used, in addition to this with dual core CPU's on the market more applications can be run at the same time which also requires more RAM obviously. I don't feel there is a big need for more performance when you consider most aps arent anywhere near being optimized with threading. I know i have never used 100% of both of the cores in my 4400+ even with visual studio 03' open , while playing world of warcraft(haha),and hosting a voice chat on x fire , with AIM open , 5 or so browsers and rhapsody playin music
 
There's also the question of "Does all this extra bandwidth make sense? Is it needed?" and the answer is yes/no. No, it is not needed in Desktop computers, Yes it is needed in high performance Servers/Workstations.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
 
indeed i do not think people realize when they are ranting about how pc3200 performs the same as ddr2 that all that extra bandwidth is not being used it's getting close but its still not there the only way i think a desktop pc could every use ddr2 bandwidth needs would be with a gaming system a bit down the road with a dual graphics card solution a top notch sound card like x-fi
and a physics card but for the average user and even the casual gamer i think ddr vs ddr2 isent going to yeild any performance differences for quite some time
 
There is also 1 thing to note about DDR2 in an Intel platform, the Front Side Bus, given it's 800MHz speed, cannot fully use all the bandwidth Dual Channel DDR2/667 offers, and the 1066MHz Bus cannot use all the bandwidth of DDR2/800, so it will be important to note about your next P4 or Conroe buy. The Athlon 64 AM2's should not have this issue, because they have direct access back to the CPU.

EDIT: Talking about bandwidth, mobo manufacturers need to remove that obsolete PCI slot and replace w/ PCI-X, most people don't realize though, if you have a sound card, 1Gbps NIC Card, and a PCI - IDE/SATA card, or any combination, you aren't getting full speed out of your components. Your NIC alone can take up to 85MB/s of the PCI Bus Bandwidth (133MB/s for 32-bit 33MHz PCI), throw in a IDE/SATA card, and you're askin' for low performance. Each PCI-X slot has a dedicated bus of 2.1GB/s (PCI-X 2.0 266MHz), much better for future things such as 10GbE and future expansion cards (or a myriad of HDD's in a RAID).

~~Mad Mod Mike, pimpin' the world 1 rig at a time
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom