Quickly. Check this out.
This is what many have been using to suggest AMD chipset has SATA/IDE problems:
http://www.techreport.com/articles.x/15256/9
Comment: No doubt that shows abysmal perf on AMD SB750 chipset. Is it true though or has Scott errored like he has many times strangely since Phenom release?
Another review debunked that with better testing (usually I am the last to quote this place, but they have improved accuracy):
http://enthusiast.hardocp.com/articl...50aHVzaWFzdA==
Comments: They both show results opposite to one another! Whats new, this is increasingly common nowadays with reviews.
...Shall we see what I get with
slower casual SATA II drives that have 3 OSes installed each, with NCQ/AHCI disabled, in IDE mode, using SB700 and HDTach 3.04 8MB test using
X2 4850e/HD 3200/ASR 780G/WD Caviar 80GB 8MB cache and Caviar SE 250GB 16MB cache?
The results are in-line with what [H] showed but totally against what TR showed. My very-slow bashed-and-smashed 80GB 8MB cache drive which makes dying noises just spinning is faster than their Raptor X 150GB. Gee.
Quote:
Originally Posted by justapost
Well today I had to clear the cmos and boot without hd's and cdrom's first to load a saved profile to get the system up. So it seems that is getting worse here too.
|
Damn. They're all buggy. None I'm seeing without problems. Not one model!
You know the most pathetic thing? The FXCN 790FX
only after trial and error my friend found out it doesn't boot nor support the 9950BE! Official FXCN techs told him this.
Quote:
|
So I still can not get a benefit out of this new board in terms of frequencies.
|
IIRC, you could only bench 3100 4core before on water, right?
This time you've benched 3400?
Your tops before was 3G 1.375v, right?
And now you can get 3.1G 1.475v stable?
=====
Anyway, 45nm rumors which I ignored actually seem to be true after more info I received. It does seem AMD has delayed 45nm desktop to January, Propus to before March and triple cores to April/May.
Facts
2nd September more details and launches come.
5050e is 45W 2.6G
8550 2.2G, 8450e 2.1G, 8250e 1.9G, Kuma BE 2.3G come now - Kuma is not native 2core but 4core disabled.
8850 2.5G comes October
8850 2.5G BE comes December
9850 95W comes mid-Nov/Dec
75% sure Rumor
DT Deneb at 125W 2.8G and 3G in Jan. Not FX... sub $300.
FX 1P will be 1 notch and 2 notches higher in multi.
50% sure Rumor
Shanghai pushed back from October to near Xmas now (mid-Nov), like Barcy was.
Opinions
If the latter 2 are true, AMD is having problems. Delays are always problems, simple as that. AND.. I won't hasten to add, if those delays are real, there will be more delays coming and the FX line will probably never appear. If they do screw up like that, they can expect to be lower by end of 2009 in debt/losses and with far less support than they have ever been. I know their business plan. They want to capture server and are ramping volumes in Fab 36 with help from Fab 38 to get as many in server segment before Nehalem can launch 2P/4P/8P. They know Nehalem doesn't have the volume to launch their yet and are trying to capture the highest profit segment before Nehalem and Dunnington can. They don't have capacity to produce in high volume for server and desktop together so they're gambling to one side and postponing the side where they have far tighter margins and lower possible profits (desktop). They don't understand that Xmas is the max spending/profit period for general customers whereafter prices drop (which they'll have even more trouble competing in). If there is no Deneb DT selling by Xmas, virtually no one will buy a Phenom 65nm, unlike now and only a few K8 will sell. By the time they have something then on DT, around March, Nehalem 2/4 core will be shipping and they'll lose out even more. Intels 45nm will cover top to bottom filling gaps that are there now. I can imagine them release 20/30W CPUs then too cutting out AMD bottom end sales and capturing over 92% of the DT market. This sort of planning is a throat chopper, AMD should know. All based on IF.