The New AMD Building Thread

Soph,
I didn't mean it that way. I was just pointing out that the 9650 was the 9600 model without the TLB error. very few people ever had problems with the "Dreaded" TLB error! I didn't mean it to sound like it was an issue for everybody. In fact I've yet to see any reviewers that even came across the problem in their tests! In truth, the 9600 with B2 stepping is about 3% faster than the 9650 is with the B3 stepping. Not as bad a hit as the Software fix, but still a loss!

Best Regards,
Russ

 

I know you meant that the 4850e was a 45w chip!!!

 

That's Okay Russ, my post wasn't generally meant for you. You were reflecting a lot of the feelings that surfaced on more than one thread when I posted the info on the 9600. Since I posted the info I've had members contest and offer their usual Intel solutions and they use the TLB bug as their trump card. Then along comes the inexperienced who listened to them and a good deal floated out of their sights. I wonder how many of these same individuals know that Intel's also have errata in their processors? There was a burst of information about that over a year ago and then it got quashed, probably by Intel.

So here it is:

http://www.dailytech.com/article.aspx?newsid=7897

http://www.theregister.co.uk/2007/06/28/core_2_duo_errata/


By starting this thread you were in effect championing AMD's right to some attention for their technology, and I support that because it would be a gray day if Intel stood alone as the only processors manufacturer. Just imagine if AMD never unseated the P4 series? We would all still be trying to overclock a Prescott single core variant. Do a little internet research on Intel errata and you will find lots of info that will surprise many. HLG

 

Yup! LOL!! Post corrected, and thanks!

Russ

 

It is nice to see some interest in AMD's new technology, and I have to say that I've never come away from an AMD build without being impressed. The 780G boards are real eye openers for very little money. There's more to speed than just synthetic benchmarks. I have no reservations about getting rid of my C2D and replacing it with a 4850e and a 780G MB. It will do everything I want and do it as good or faster, and that's all I want or need! Just to see folks getting excited about AMD again (and deservedly so) makes this thread worthwhile. To be able to do it without being harassed and beat over the head with Intel this and Intel that, makes it more exciting.

For me, after putting OxiMoron together and using the 4800+ for a while, I wanted one. After the 4400+ on the Biostar 780G MB, that want turned to "Gotta Have One"! For a 2.3GHz 64x2 at stock 2.3GHz to get my attention the way it did, and to actually prefer it over my E6750, I just have to have a faster one for myself! LOL!! I spent 3 days of constant use and abuse with that 4400+ without a single hiccup or glitch just added to that feeling. Not needing a ton of fans to keep things cool was a big plus too!

Thanks for joining our little party,
Russ

 

I've known the Intel CPUs have errata for a while now, but overlooked it because I've never experienced it. The same only seems fair with the AMDs, it wasn't immediately obvious to me that the situation can't happen outside of virtualisation, I assumed it was just more likely to come up in this scenario. I stand corrected.

 

has anyone heard of this problem with the AMD dual core that somtime on some of the CPUS in the dual core range that the temps can be miss read from the cpu. is there any fix for this i have updated my bios on MB but still NO! change.

 

That can happen on any processor where the thermal sensor has gone bad or gets stuck. It's a common issue with the new Intel core 2's. My E8400 had a stuck sensor and my Q9650 has two that disagree. You have two choices for measuring temperatures, you have the core sensor which measures from each core for T-jmax reading and your T-case sensor which takes an outer measure. If you're concerned that one is wrong then use the other because it's rare that they both go bad. The most accurate measure is the core sensor but the T-case measure can help you estimate the core temperature quite close by adding another 7c to 10c to it.

 

IIRC, isn't this only a problem with core temperature readings, not the main CPU reading?

 

It can happen with all three sensors on a dual core. It's just more likely to happen with the core sensors because of their hotter environment and the more radical temperature changes they experience in use. It is far less likely to occur with the CPU sensor because of the lack of such radical change in temperatures, because it's range in use is almost always much narrower.

That's exactly what happened when the pin broke on my CPU cooler. I had a 4-5C increase in CPU temperature, but the cores got much hotter! It made my computer lock up and freeze!

Best Regards,
Russ

 

Sam

Isn't it going on about 4 AM where you are? LOL

Thermal sensors on multicore processors suck if one expects that they will all be balanced. My rule of thumb is to go with the one that shows the highest temperatures. If the core sensors don't seem to be working then I use T-case measures and add about 10C to closely estimate core temp. I do that on system that have good core sensors and it's reasonable close.

 

Hehe, yes it was
That makes sense, I tend to avoid paying too much attention to the core temperatures as so many people seem to moan about them not being accurate. As far as I'm concerned, IIRC the shutdown threshold for CPUs is the main temperature not the cores, so that's what I keep my eye on.

 

Any insight on this Monitor
DCLCD DCL20A+ 20" Widescreen LCD Monitor - 1000:1, 1680 x 1050 (WSXGA+), 2ms, VGA, DVI, Black
Original Price: $189.99
Instant Savings: $60.00
Price: $129.99
http://www.tigerdirect.com/applicat...?EdpNo=4108000&csid=ITD&body=MAIN#detailspecs

 

Not bad, but I've never heard of DCLCD. Sceptre seem to be a relatively trusted brand, and their 20"s are quite cheap:
http://www.newegg.com/Product/Product.aspx?Item=N82E16824112013

 

Ya ! But 20bucks difference

 

ROFL... It is a Sceptre Sammy. Scroll down the page rick provided and see who "provides" the warranty.

....gm

rick I can't believe the prices of some items now. I picked up my little 19" a few months ago for $165 and now they have "name" brand stuff the same size or bigger for less than $130.

 

Ah, my mistake, probably should have spotted that.

 

Rick,
If you buy a 20.1, get this one for $159.
http://www.newegg.com/Product/Product.aspx?Item=N82E16824112013
I have one and after more than 2 1/2 years of constant use (15-17 hours a day), the driver board for the backlight is going out. Flat fee of $75 to repair. If you look around the net for repair prices, it's a bargain. No complaints at all as I give a monitor quite a bit of use. Far more than average! They are close enough to me that i can drop it off and pick it up and save the shipping!

For the money difference and 22", this one for $10 more.
http://www.newegg.com/Product/Product.aspx?Item=N82E16824112015

They had a real nice 24" for $199, but sold out yesterday (Ya think! LOL).
http://www.sceptre.com/Products/LCD/Specifications/spec_X24wg-Naga.htm
One of the better black Friday Specials.

I've looked at all three and they are outstanding quality for the price.

Best Regards,
Russ

 

Remember I'm building and getting this for my Grand Daughter. I'm not cheap, but I have to watch my budget

 

greeny and sammy take care because you guys both know that Russ has a 20.1/22 inch Scepter LCD.


My wife purchased the 24 inch scepter because it was going for a low price. I don't care for anything with TN panels especially for Video playback, but hers doesn't look too bad when playing 1080P. I prefer S-PVA or S-IPS panels because of their superior contrast and color. Both Panels have evolved over time and there is a debate among which is better but if you get either one of those panels you will be happy.