Jump to content

Well, it's the only source of info...


Recommended Posts

Considering the source, it wasn't a too terribly written preview of the upcoming platform.


But, as with any preview, all bench results should be taken with a grain of salt. The final mass-produced product could perform much differently.


As a personal preference, I'm sticking with socket 939 until I see the new platform prove itself.

Share this post

Link to post
Share on other sites

The results showed absolutely no performance gain going from DDR400 to DDR2-667; therefore, AMD is going to wait to support DDR2-800.


What I want to see is how the current batch of DDR500 thru DDR600 compares to DDR2-800 or even DDR2-1000. I'll bet money that Mushkin Redline or OCZ Plat EB PC 4000 blows the pants off the DDR2.

Share this post

Link to post
Share on other sites

I did some research with my AMD rep today and she said the bosses were none too happy with THG. AMD required in writing that they not release any benchmark figures since they were dealing with early engineering examples. What THG got was basically a technology demonstrator. The silicon has no resemblance to a production processor.


This line is vintage THG...


"Whether the current engineering sample processors actually suffer from a memory controller bug is hard to say. This could also be due to information that was released intentionally to prevent people from doing early benchmarking."


Speculation based upon speculation is not the trademark of a conscientious organization.


But on the very next page they slam anyone that would speculate on the performance by saying this...


"We do not want to speculate over the existence of a bug in current silicon as reported on the web, since the processor we used is too far away from mass-production anyway."


Well guys, you can't have it both ways.


Another pitiful example of unethical reporting by THG?


You decide!!!

Share this post

Link to post
Share on other sites

I agree, THG is a bit, well strange.


His reviews are offen one sidded.

Not explorering all the angles.

Sometimes dissing a product just because it does'nt perform in one area, then not explorering it to it's fullest, or even explaining why it was failing.

Ie the psu reviews.


Also his lack of knowledge in certain areas, is well, not great.

Ie his memory reviews.


His reviews really are'nt that bad, about as good as any other out there.

Sometimes with a bit extra in there.

Sometimes not.


You cannot truly judge a product by it's reviews that's for sure, ie video cards and psu's, and even memory.

Also mobos are the same, no review is gonna tell you everything about it.

Especially the video cards, man I will never look at another review for a video card, they were totally off.


Some reviews are screwed, ie hand picked products, the likes the consumer never see's(memory, ie - you pobably know what I'm talking about).



Also, when signing an nda, it's best to adhere to that nda if it pertains to your future work ^^.


As for the platform.

Sounds great for the oem and the majority of people out there.

DDR2 actually can be allright I guess, some with halfway allright timings.

The speed is the good thing for the oems and such, ease of use sort of thing.


Might be worth it down the road.

I myself think whenever I do upgrade, I'll be going for the nf4 expert board.

With dual optys, with some samsung memory.


If they were making a new platform with gddr3 or gddr4, I'd be all for it.

That stuff is dang cheap, under a $100 for 512m, for the oems I guess, not like we would get it that cheap.

1200mhz min(as long as the timings are right, otherwise more like 1400mhz give or take, sometimes more) with "samsung" gddr3.

Over 40k a sec with 256bit interfaces...


4 sticks, 64bit a piece, upgrade the memory controller for something like that and the memory type...

Othrwise bandinterleaving may help, 128bit you can get over 20k with.

Or make it so it's 128bit per stick...., 4 banks per stick, just make some new slots...

There's I'm sure many ways to put that together, with either bankswitching or with direct access.


That's double the top memory bandwith today on a nf4.

Even with the same bit width.

To late now I guess, damn shame no one wants to really move forward, at least the oems have there day lol, ie dell.

You're gonna have a fast dell dude..., aww...


I'm really sick of seeing the cpu's stuck under 5ghz for amd too...

We need that speed allready, come on amd, it's been years since someone has reached 3ghz.


Even though I can't run that high, it's not like I dont' need it...

Even if I could run 3ghz, it's not 4ghz or 5ghz which we realy need.

We need 5ghz by the end of 2007 ^^, I'm not joking....

Nor am I being retarded.



Anyways since everyone and there mom will likely move on to the next socket.

Likely these older sockets will get somewhat cheaper, maybe...

Then again, maybe 939 and ddr willl die off and make things more expensive!, that does'nt sound to good...

I still need a cpu, and memory, and mobo...

Share this post

Link to post
Share on other sites

I was not even going to read that article untill someone else reported on it. It sounds a lot like they have ben reading other tec sites and not restating without a true understanding of what they are talking about. Complaining about a bug on a early preproduction CPU is stupid. They dont do that to intel.


How much Intel money do they get anyway?


The only really nice thing about the new socket that I see is the new for hole mounting which should allow better mounting without a IHS.


Was AMD reporting thaty they were going to quad core on this socket? It really doesnt look like much of a change.



BTW these were good results for such a early production chip especaly considering Toms very limited tweaking abilitys.

Share this post

Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in

Sign In Now
  • Create New...