Quote:
Originally Posted by Tony
[Van diesel]I live my life 8 seconds at a time [/Van Diesel]
Thing is 1M superpi is now at the stage that it's not really proving very much, and is in general residing as purely a benchmark with an ever waning real world reference. I think we're gonna see a divide on opinions here, those that want a system for "real world applications" likes games and those who are in it for the benchmarks.
All in all we are being spoiled for choice if you ask me
They OC well, so what? A celeron is still a celeron - crap ;)Quote:
Originally Posted by Hornet331
The cache is lacking on them. And before you say "yeah well the celeron 4xx has the same L2 as many K8 chips" remember that AMD's architecture does not rely on cache much at all, just doesn't need it near as much. Once you have 512K then any more the performance gains are minimal for most applications. This has been proven when comparing two K8's with the same operating frequency but one had 1MB L2. Intel on the other hand is much more cache dependent so it has a greater performance hit with a lack of cache. I'm no CPU architecture guru - far from it - but I'm thinking part of the reason is due to the lack of an IMC and direct connection to the memory like K8/K10.
For the same price as a Celeron 440 I can get a Brisbane X2 4000+ 2.1GHz. Plus the boards are cheaper for a decent one. Not a hard choice here.
Thing about Super-Pi is that some guys have said on the upcoming of the Barchie, that stepping B1 or BA did 14sec at 3GHz.......THAT'S why so many ppl are confused.....They are trying to find out from WHERE this 14sec came out and from WHERE this 30k of 3DMark2006 came out..... ;)
Propably AMD in their labs, had a Yorkfield and after some changings here and there, they though that they were benching with a Barchie or a Phenom and got that numbers.... :D :p:
PWN Deluxe :up: :clap:Quote:
Originally Posted by Vapor
Anyways does SOLDNER-MOFO have anything to say about these SPI scores? did he not see some crazy score form a friend of his?
maybe it's time for amd to take the handchief on the phenom logo to dry their tears :rofl:
No......It's time to be honest and come out and say: "We have this, we sell it that much and that's it"......Don't put me snakes that will unleach their power and bite you.....Don't put me dogs/cats/etc annimals........Put me a picture of the Phenom and say: "That's it guys....if you want it, get it"......:)
Someone banned him, for what reason I have no idea... Maybe because he had something good to say about AMD? :shrug: (I KID! :p:) I don't know if it is a temp vacation or permanent, heck I don't even know what happened, just logged on one evening and saw "banned" under his username :shrug: So I don't think we'll be hearing from him for awhile at least.Quote:
Originally Posted by [XC] riptide
hmm not where i live. :p:Quote:
Originally Posted by SparkyJJO
cheapest X2 4000+ 50€
cheapest cleron 27€.... and for 31€ i get a board...
on the other hand a Pentium E costs also around 51€.
you should have compared it to a sempron, the cheapest i can buy costs 24€.
anyways this goes far to offtopic.
well if you go by cheapest celeron those are definitely less than the X2 4000+ but it also isn't the 4xx series either that you mentioned, so....Quote:
Originally Posted by Hornet331
But yes, definitely off topic now...... :yawn:
OK I need to get my homework done but I'm keep getting distracted... somebody temp ban me until 4 PM EDT today so I can get it done... :lol2:
I already expect on the stats that Phenom BA/B2 will lose over Penryn in;
Multimedia
Overclocking
Super Pi (not that I give a damn below 32M although a good extra addition like gold laces on my track shoes :D)
At same MHz.
The only thing that matters to me is real world daily applications which MOST desktop users run at one time or another since that is exactly my interest. Videoing, 3Dmodelling, Imaging, Gaming, Compression, Decompression, Encryption, Decryption, PCMark 04/05, Cache2Cache, Science apps, SYSMark and EVEREST Ultimate benchmark suite as an addition. Sandra, MemTach, CPUMark, CrystalMark, ScienceMark, RealStorm2006, Cinebench9, Cinebench10, AquaMark, 3DMark 01SE/03/05/06, POVRAY, Vecmat, wPrime, CPU-Z latency, Prime95 will also be very helpful for additional info. Also to see power draw at 2.5GHz and 3GHz, volts needed, temps, speed scaling, headroom, coldbug and retail price.
With any new product you expect improvements. Core 2 over P4 was incredible. People are now stuck in a trance expecting the same each time which is never true but for luck. Penryn is just gaming/multimedia optimized die shrink Core 2 so improved. Pentium 4 631 takes 6.15GHz to do 21 second 1M. Q6600 2.4GHz stock everything with stock DDR2-800 does 21 second 1M. A huge change. To beat that by AMD would be a huge improvement (in what?) like Core over P4 but it means toot to me if done or not.
I agree.Quote:
Originally Posted by Vapor
I guess.... P4 ;)Quote:
Originally Posted by BeardyMan
3DM03 + poor vga ;)
C2D E6300 @2.8GHz, 2 cores, 1.3vcore AC
GT1 - Wings of Fury 195.3 fps
GT2 - Battle of Proxycon 29.1 fps
GT3 - Troll's Lair 27.6 fps
GT4 - Mother Nature 40.8 fps
P4 631 @4.8GHz, HT, 1.3vcore AC
GT1 - Wings of Fury 190.1 fps
GT2 - Battle of Proxycon 29.2 fps
GT3 - Troll's Lair 27.6 fps
GT4 - Mother Nature 41.0 fps
Same score with aircooled K8 is IMPOSSIBLE :D
hipro,
I know you are an xip and all, no dissrespect intended, but you seem to be the ONLY xip on xs who is so in love with Intel. Thier may be others but they keep thier bias to themselves. Why get involved with all the petty bickering?
Aircooled K8's are running at 3.4-3.5ghz now :DQuote:
Originally Posted by MaSell
totally agree with you hipro.Quote:
Originally Posted by hipro5
For me it's simple... just post a 3dmark2001 with same gfx and OS.
That should say all to me...
Fixed. But to be fair, whether the benchmark sucks or not, it does give us some kind of standarized indication to talk numbers, and in the end, that is what drives this industry, not some kind of hocus pocus methodology that says this product is better without fact. If there is a benchmark, a company will try to get an advantage in it, or based on its compiler or architecture, will be optimized for it. With the case of SuperPi, everyone knows Intel chips does much better than their AMD counterparts.Quote:
Originally Posted by hipro5
Yes... nowQuote:
Originally Posted by BeardyMan
But we talking about 2yo stuff :D
Actually, I recall a certain Pentium-M owning both in Super Pi, and also being pretty good for games (if not as good for encoding).Quote:
Originally Posted by BeardyMan
More to do with being able to fit everything into its very fast cache, though.
Edit: so, to beat Intel at Pi, AMD would probably need to cut their cache latency in half and then some. But then again, AMD isn't so dependent on cache because they have pretty good latency to the RAM.
Wow... that 32m is slow ?! I'm sure i get sub 25min with k8 @ 3ghz and rams 250/2-2-2 :shrug:
I know that phenom has 512k versus 1mb of the a64 3700 but i was pretty darn sure that it would keep up cpc =/
yes, pi ain't everything but it's a good measurement of raw cpu/ram/mobo combo speed , imo .
fixedQuote:
Originally Posted by Hornet331
Tony, how do you really feel about AMD?
Heh, I should start a new super pi database.
I only care to see real time gaming application performance numbers.
Dissapointing results.
Whammy! :hammer:Quote:
Originally Posted by Vapor
Really hoping that vcore isn't correct on CPU-Z. That 32M time really blows though, I'm pretty certain if I had a calculator and a sheet of paper I could calculate it faster. Wonder if it'll be possible to break 3.4GHz with Phenom and ~1.65vcore and some a non-coldbug tuned singlestage.
Good Sandiego chip on air+BH5@DDR570+ would do the trick.Quote:
Originally Posted by MaSell