The Official PC building thread -3rd Edition

Theone: That's AMD's HD series graphics rendering at work, you can see why I recommend them so much! I know exactly what you mean by saying that the colours are 'lush' but it's nothing you can really get across without someone having seen it themselves.

Rick: The HD3200 beats the GMA 3100 hands down, it's triple the speed for 3D rendering, provides a much better picture, and usually comes with more outputs. AMD's integrated efforts have left Intel standing.

Abuzar: Faster system, but worse graphics chip, depends on the use I guess. Actual performance wise (outside games and HD video), the Intel system you listed would be around 30-35% faster, when clocked to 3Ghz, more like 60%, but I'd have to stick an HD3450 in that system to get it to be as good at everything else as Russ' AMD, and that will push the cost up a little bit.

Rob: Great news! Nice to see where your spare funds end up! I'd strongly advise going with a Corsair TX unit over an OCZ, especially one like a GameXStream or ProXStream. The system will probably run fine without the extra 140mm fans, but if you get heat issues, you at least have the option!

 

yea i think im gonna stick with the corsair tx650w. ive heard alot of bad thiungs about that OCZ PSU.
thinking also going with a HD4850 maybe this one
http://www.newegg.com/Product/Product.aspx?Item=N82E16814102770
alot better than the 9800GT.which is soon to be obsolete.

 

Yes indeedy, the HD4850 is a good 20% above the 9800GT, perhaps more. I've read nothing but good reviews about that particular sapphire card, and it's my card of choice for system build specs at the moment.

 

With the launch of its latest quad-core Opteron processor, code-named Shanghai, on Thursday, Advanced Micro Devices is looking to rebuild confidence among customers after problems with its Barcelona chip earlier this year. Running at 75 watts, the Shanghai processors range in speed from 2.3GHz to 2.7GHz. The chip also includes improved hyperthreading for faster application performance. Manufactured using the 45-nanometer process, Shanghai chips are more power-efficient than Barcelona, which were manufactured using the 65-nanometer process. New CPU designs allow Shanghai to turn off unused functions or shift processing duties to save power. It also provides more power savings when servers are idle compared to earlier processors.

hear is where I read this
http://www.rlslog.net/amd-ships-new-quad-core-cpu-shanghai/

 

A 75W Quad core. Now all AMD need to do is get some overclocking abilities

 

sammorris..I thought AMD were good overclockers

 

Not at all. The later Athlons are the worst of all, to squeeze 20% out of them was a pretty impressive achievement, to not achieve 40% out of a Core 2 almost seems criminal...

 

Well how are the new chips (AMD)a Stock settings
I'm really impressed with the Intel 2.4 & 2.6 Quads, they are really fast at stock (no OCing)

 

I can tell you that judging by the performance of the AMd build I just built. AMD is slowly getting there. To be able to make a 4400 Dual Core perform almost on a par with a 4800+, is quite a feat!

BTW, we were talking about LCD Monitors for Medical use the other day. How about a Black and white monitor for Mammography and Radiology! 20.1" 4:3, with a resolution of 2048 x 2560 and a Pixel Pitch value of 0.156mm. From NEC on sale for $7642.76! Wow!

Russ

 

Almost? Don't mean to sound critical, but surely a 4400+ was always 'almost' as fast as a 4800? There's less than 10% in it!

 

Only if it was overclocked. Stock it was about 20% slower!

Russ

 

hows it going fella's. looks like things are looking up for me. JOB+
COmputer+
think i got all the buggs worked out now. the problem i was haven with my HDD's sata header or cables what ever it may have been,
i fooled around with the cables got 2 of them tight in the header and no errors for hte past 24hrs. replaced 2 others about 6 hours ago, no errors. so got my new mushkin mem in and running fine. all is well.

GM
the problem i had this morning with the sound beeping like i was unpluggin a usb cable or something. fixed it was well..lol
good things come to those who wait.

 

Congratulations Rob! Glad to see you found somethin, obviously better for more Money.
I have a question. Do you fold your Sata cables for neatness? The reason I ask is because I read in Tom's Hardware a while back that doing this can cause problems. I've been using Sata drives since Sata 1.5 on my D-940, and have never had a cable failure. I've accidentally knocked one loose from time to time, but never had an Actual failure. Just a thought!

Russ

 

Er, what?
Read the figures again Russ...
The 939 4800+ vs the 939 4400+ was a 9% difference. I don't see why that would have changed with AM2.

For the record, I've had several S-ATA cable failures from bending them flat. Don't do it folks!

 

Yeah, according to the AMD performance ratings. Are they really accurate though?

 

They are actually. The 4400+ and 4800+ figures happen to be exactly double the processors' clock speeds. Since they are identical (same architecture, same cache), the difference is literally a mathematical division of the two numbers, which makes 9%. I'm pretty certain this was confirmed in the Sandra benchmarks.

 

My rule of Thumb is to Roll my Cables as small as I can & use a tie to hold them in place and as far out of the way as possible

 

Different cores, with half the L2 cache compared to the socket 939 Windsor. 2x1MB vs 2x512Kb The socket 939 chips were better performers and better overclockers, particularly the 4400+, which could almost equal the 4800+ in performance when overclocked. If you remember, the 4800+ wasn't worth the extra money because the 4400+ OC'd so well! I know a couple of years ago I was thinking about getting a 4800+, and I think it might have been Sophocles who pointed out that the 4400+ 939 Windsor was almost the equal of the more expensive 4800+. The newer Windsors are no match for the old Socket 939s, and don't OC all that well!

Russ

 

the 4400 had 1mb, it was the 4200+ that had 512k. Not that it's that important.

 

Please provide a link and some empirical data for such statements. I've been doing this for over a year now and NO evident problems that I can tell. What are the problems that they indicate? It's truly not that big of a deal because I can loop them up and not bend them but folding them is so fun.

....gm

add: RedRob good to hear you got it sorted. What was it? Loose connection? Gremlin? or WHAT??