I dunno if these can be trusted, but someone else at a different forum showed these to me and I thought you guys might like to see them
http://forum.coolaler.com/showthread...=158813&page=1
What do you guys think
Printable View
I dunno if these can be trusted, but someone else at a different forum showed these to me and I thought you guys might like to see them
http://forum.coolaler.com/showthread...=158813&page=1
What do you guys think
:BS:
CPU-Z wouldn't recognize the CPUs yet... not true...
that's my train of thought, the only problem with that is coolaler does have a reputation of getting es chips before anyone else, but the cpuzs seem fake to me
CPUZ picks the chips up and they are in the wild.
perhaps, but we know for a fact that there are 45nm es chips out there because mobo manufacturers have begin designing bioses for them, here's asus's list of supported mobos
http://event.asus.com/mb/45nm/
The screenshots are likely legitimate. The latest version of CPU-Z does in fact read Yorkfield correctly. I've seen other similar screenshots.
Those are legit.
I added a poll, this will make it clear what everyone thinks of them
They said it needed an upgrade to recognize them correctly (at least I think..Quote:
Originally Posted by Raptor22
some posts point to that)
http://www.cpuid.com/cpuz.php
1.40.5 07/06/2007
* NVIDIA nForce 560 chipset support (1.40.5)
* Intel Dynamic FSB technology support (1.40.5)
* Athlon 64 X2 BE & Sempron LE (TDP 45W) (1.40.5)
* SiS671/FX/DX/MX northbridge support (1.40.5)
* SiS968 southbridge support (1.40.5)
* VIA P4M800CE chipset support.
* Preliminary support for Intel Penryn CPUs family (45 nm).
* Intel Core 2 Duo E6x20, Pentium E2140/2160 processors recognition.
* Intel P35, G33, G31, Q35, Q33, X38 chipsets support.
* Intel GL960/GM965/PM965 chipsets support.
* AMD Athlon 64 "Lima" processor support.
* AMD Geode LX processor support.
* NVIDIA nForce 520 chipset support.
* New parameter "-console" to generate output in a command prompt (Windows XP only).
* New icon, courtesy of Techlogica.
looks like these might actually be legit!
I think these are legit.
Just wondering baron, you wouldn't happen to be baron matrix at toms hardware? If you are, I find that kinda ironic, if not, then it doesn't matter
2.33 is in line with the cpus tested by MSI. What score does a Core2 get in 3dmark06 @ 2.33?
only a 7x multiplier... That's kinda restrictive.
Ryan
yeah, but supposedly x38 will make up for that, I mean p35's already hit 672fsb (victor wang and oskur wu did it), it's only so long before we see 700fsb, which would mean 4.9ghz which is pretty nice for a low end chip
well if that's the case I'll shut up. I wasn't aware the p35 would go that far. Is that a DFI thing? How far would another brand go, say, Abit?
Ryan
I hear the i35 pro is supposed to be very good too, but for once asus seems to have gotten it almost right with the p5k series, I would also recommend those boards too (just not anywhere near dfi's). I expect to start seeing a few good things from foxconn too. Why? Because they got a crap load of dfi's engineers within the last year, so hopefully things will improve for them
here's the wr
http://www.xtremesystems.org/forums/...d.php?t=152099
Actually I believe this is the current record.;)Quote:
Originally Posted by AliG
http://valid.x86-secret.com/show_oc.php?id=220153
When looking at CPU-Z v1.40.5 I found this, so, I'm pretty sure it does detect it.Quote:
Originally Posted by Raptor22
http://img510.imageshack.us/img510/8...olfdalenf8.jpg
out of all 12 pages of that thread here's the best pics
http://www.coolaler.com.tw/toppc/45NM/500PI.JPG
http://www.coolaler.com.tw/toppc/45NM/50006.JPG
http://www.coolaler.com.tw/toppc/45NM/540PI.JPG
3.7 @ 1.1v. Wow.
Ryan
http://www.hkepc.com/bbs/hwdb.php?ti...iew&rid=837360
Wolfdale benchmarks. Nothing more there to argue about. Also matches Intels early preview show :p:
I doubt that is a correct Vcore reported in CPUz.I don't say it is not possible,i just doubt it :).
what's with the 300 & teletubbies stuff?
AMD better have their crap together w/ K10 b/c from the HKEPC link it is obvious that Intel is at the top of their game.
45nm FTW baby!
Looks legit.
So what are the real world benefits of Wolfsdale and Yorkfield? Besides the 45nm manuf. process and larger cache. Don't see any temp readings in those screens. They should run at lower vcore and give off less heat right?
edit: nevermind, found the hkepc link. :rolleyes:
http://www.hkepc.com/hwdb/intel-penr...view/epfpi.png
Here's another one
http://www.hkepc.com/bbs/hwdb.php?ti...iew&rid=837360
results look pretty good, especially power and temps, 37 load is pretty damn amazing compared to a 49 load
edit:whoops, looks like someone else beat me to it
Load is the same as normal body temperature haha :D
http://biomed.brown.edu/Courses/BI10...s/ice_cube.jpg
Yeah, that looks great. Especially the games and power/temp results. :)Quote:
Originally Posted by AliG
Idle (EIST Disable)
Intel Core 2 Duo E6550
Power - 61W
Temp - 41C
Intel Wolfdate 2.33GHz ES
Power - 43W
Temp - 31C
Max Power Loading Tools for Vista
Intel Core 2 Duo E6550
Power - 83W
Temp - 49C
Intel Wolfdate 2.33GHz ES
Power - 59W
Temp - 37C
:yepp:
does yorkfioeld/wolf have different arch. or new instructions or somesuch vs kentsfield/c2d?
Larger cache, a few optimizations to the pipeline (but not enough to be called a true revision to the architecture), and SSE4 instructions (hence why encoding apps got huge boosts).
IIRC, On Anandtech there is also supposed to be a way on these 45nm chips managing the load on the cores so that one core gets clocked up if only that core is in use or something.
Thats for laptops only.Quote:
Originally Posted by direowls
Could you show me a link of that? Didn't say that in the Anandtech article...Still, it would explain why the pi scores are pretty much the same as C2D's, despite pi's focus on bandwidth.
Personally, I thought this was pretty cool:
http://img.coolaler.com.tw/images/mz...gjrmmn4wwo.png
472 FSB ain't bad ta'll...and we're talking a MSI board here...
It's also within the decent strap, right? IIRC it changes at 490 or so?
Hmm. Perhaps the quad to get this time will be the 8x multi chip (X3330/Q????). :D
http://www.hkepc.com/bbs/hwdb.php?ti...iew&rid=837360
http://www.hkepc.com/hwdb/intel-penr...iew/edat-2.png
Not sure if it's still mobile only though it is precluded with some info on new mobile sleep state.
Does anyone know the pricing of Wolfdale/Yorkfield yet?
MADNESS!Quote:
Originally Posted by squilliam
Like these chips :yepp:
Wondering about their pricing and ETA...
No, but it would be my guess that they will line up with their Xeon 771 Brethren, although there probably are less products:
X5460 3.16 GHz $1,172
E5450 3.00 GHz $851
E5440 2.83 GHz $690
E5430 2.66 GHz $455
E5420 2.50 GHz $316
E5410 2.33 GHz $256
E5405 2.xx GHz $209
If there is a 775 (x33x0 or whatever desktop Yorkfields are called) version of the E5430, I'm all over that. :)
8.5 multi for ~$300, yeah, that'd be yummy. I'd settle for an 8 though. :)
As for ETA, I think we're still looking at Q407 for EE, Q108 for wolfdale/yorkfield...Don't think they got more specific than that. I imagine we'll have to hear something coming out of IDF SF.
yeah, but you have to take in mind that's on only 1.08v, so 3.3ghz is pretty nice considering what temps you'll be getting compared to a current quad at 3.33ghz
I personally would just stick with the e5410 and use the money on a better mobo (I'm not a fan of half mults)
Thanks red. Looks like it is connected with that deep sleep state tech.Quote:
Originally Posted by red
Quote:
Originally Posted by AliG
Oh, it definately is nice, but I doubt even with more volts the FSB will go much higher than that on regular cooling (I'm a simple water guy myself). So, to me, the only big question mark left is how high can the frequency scale, as it should be able to hit around the top FSB of the most decent strap. I figure around ~4ghz on avg (still waiting on someone that gets an ES with a higher multi than 7), so 8-8.5 multi would be the sweet spot, and considering the 8.5 is where the bang-for-buck ratio hits, and 8.5x471 = 4ghz...Yeah. Granted, if desktop (or X3300 series Xeons) only come in full multis, I think 8 is the way to go.
Granted, still waiting on more results, but these are encouraging and spawning thoughts of future upgrades. :)
woah Intel is working hard.. :) very good for us consumers.. :) better product..
hehehe hope the pricing is good like the one that we are experiencing today..
did they say that 45nm have high yields??
what I would like to see is how high dfi's x38 board will scale. If that thing can hit 600fsb out of the box (considering the quality of the board and the cooling the p35 board has, shouldn't be a problem), then even with an 8 mult you can hit nearly 5ghz. If we're lucky, with some patience and tweaking we'll be able to see a 700fsb board, but that may take a while
yeah, I don't think amd gets that in order to get sales, you have to work hard on the product, rather than making fun of your competition; and that the consumers are your friends and source of money, so you should please themQuote:
Originally Posted by erwinz
http://media.corporate-ir.net/media_...sHoltJun07.pdf erwinz, slide 20 shows that 45nm is trending as 65nm yieldwise and targeting 65nm's current yield rate.
that can only mean either amd has some huge guns in plan, or they're in deep :banana::banana::banana::banana::banana:
yup yup.. hope Amd will show some teeth.. hehehe I'm also waiting for quad core AMD.. :) I'm satisfied with my dual core conroe.. I'm not jumping at quad core, I want to see the both camps.. hope AMD can deliver to what they are saying.. :)Quote:
Originally Posted by AliG
hey that is very nice news.. :) cheep processors again.. very nice.. :) I can't remember the time I this that we have so many options on a good processor.. and still on a budget.. :) (buy a budget processor and mild overclock it.. hehehe)Quote:
Originally Posted by red
yup yup.. AMD should respond.. and fast.. :DQuote:
Originally Posted by AliG
Those aren't "clock/clock" numbers btw. Intel was comparing 3.33GHz with 1333FSB with 12MB and other goodies compared to 2.93GHz, etc, etc. You can verify the 15, 25, 40 numbers with this graph http://www.tgdaily.com/content/view/31669/135/Quote:
Originally Posted by higgins
Quote:
Originally Posted by FghtinIrshNvrDi
If I had to guess, I'd say 1.100v means the vid is 1.212v, and vcore is actually set above roughly 1.33v. How far above 1.33v, it's not possible to say. Too high? likely :p. wonder what intel's spec sheets will say the max and abs max vcore for these cpu's will be.Quote:
Originally Posted by informal
and it's good to see many p965's that will indeed support 45nm after all :)
well, we will see what amd has in a few weeks hopefully. wolfdale is looking sweet, i hope barcelona will be able to at least catch up in a few spots...
Did you ever doubt it will ? ;)Quote:
Originally Posted by Shintai
Coolalers web is trustable place for me ...
Legit or not, the pics are a temptation :slobber:
Coolaler is to be trusted, no speculation needed.
But the cpuz can't be trusted with voltages
LOLQuote:
Originally Posted by higgins
Since when a Conroe at 2.33 consumes 83w? I thought it consumes clearly under 65w...
1.40.5 under. 1.140.8 fixed the issue, so do not take the voltage serious.Quote:
Originally Posted by red
indeed Coolaler is trusted and i think to question it with a poll might insult some fox.
indeed.Quote:
Originally Posted by gillll
Hehe, they showed that to us at a BUM (business update meeting) everyone cracked up when we saw the last slideQuote:
Originally Posted by red
Quote:
Originally Posted by AliG
eeek, imagine the horribly nasty latencies at 700.
i dont see any benchmarks on colaaler, and wth is up with all those teletubbies there? :stick:
about this:
http://www.hkepc.com/hwdb/intel-penr...view/epfpi.png
40% more what? performance? performance per watt? im 100% sure that this isnt a 1:1 compare between 65nm and 45nm silicon.
its higher locked 45nm silicon with a faster fsb compared with slower, low fsb 65nm parts.
Its performance. But its also based on the frequency difference.Quote:
Originally Posted by saaya
Dunno if someone already posted it but we have a lot of benchmark here :
http://www.hardspell.com/doc/hardwar...page21462.html
Enjoy
Its HKEPCs results they use.Quote:
Originally Posted by Tobalou
115% with DivX sounds amazing. Anime FTW.
well how can you be so sure?Quote:
Originally Posted by Shintai
they dont clearly state it, and in case the performance isnt there they can say its performance per watt, which is easy since they are on 45nm.
Oh they do state it. Its just a matter of seeing all the slides and such.Quote:
Originally Posted by saaya
This should also give you a hint :p:
http://www.anandtech.com/cpuchipsets...spx?i=2972&p=3
looks like that faster 24-way cache does seem to make a difference, everyone thought it would be a waste of time (of course a lot of the improved scores could be mainly due to sse4)
Now, I don't like to use this word but you do have to be a real NOOOOOB to question coolaler's pre-NDA results..
This is months before Conroe was ever available:
http://www.xtremesystems.org/forums/...ad.php?t=95021
Don't pee-off the people that make this forum xtreme, man! :ROTF:
Dang, 480 FSB on the quad. I am impressed. I wonder how they will do on the DFI P35 or the X38's.
http://img.coolaler.com.tw/images/5u...mdmn4znq42.png
http://forum.coolaler.com/showthread...158813&page=14
http://i11.photobucket.com/albums/a1...y3mj1gtyzz.png
Yorkfield 2.33 at 3.2 doing 14.422 SPI
http://i11.photobucket.com/albums/a1...e/E67_3_45.jpg
E6700 2.66 at 3.45 doing 14.672 ambient 33C
Not bad at all.
1st SS drom link in OP
Dang!Quote:
Originally Posted by Eastcoasthandle
QMFT!Quote:
Originally Posted by -thc-(cZ)
Looking good, here is an overclock they did yesterday.
http://img.coolaler.com.tw/images/n3...0dqyv2gomt.jpg
wow that looks really nice, ~10% ipc improvement at 4ghz.
I just wonder wether the ipc improvement is lower at 3ghz?
Is it just me or does Yorkfield look a little sluggish?
Me Wants Yorkfield!! :D
Shhh :p: ,you are not supposed to say things like that,you will invite a couple of knights for truth into this topic and than we'll never see those screen shots from coolaler forum from all those "truth-spam" posts :).We will only see intel slides & nice numbers.Quote:
Originally Posted by Eastcoasthandle
So no,it doesn't feel sluggish at all :).
On the other side,does anyone who speaks Chinese knows what was the vcore for that OC with yorkfield?Hard to swallow those 1.1V in CPUz for 4Ghz.
I don't speak chinese but I actually read the coolaler forums and used google translate and btw the numbers aren't chinese :DQuote:
Originally Posted by informal
1.37v @3.36ghz QC
then there's a 4ghz pic of a DC and some talk about 1.45v later
those numbers aren't that good, but you never know how far away the used ES are from retail silicon (i.e. how many spins are in between).
I don't think that yorkfield looks sluggish at all so far, however if you expected wonders (like 20% ipc gains core 2 had...), well then maybe.
edit:
on page 8 you can see that a 3,2ghz yorkfield consumes about 46 watt less than a 3,2ghz conroe.
imo this indicates 3,3ghz is well within tdp even with those ES and 3,6ghz quads should be also doable for Intel in the near future...
on page 4 there's a 3,2ghz QC pic at 1,25v
Well, lets just say that if I plan on upgrading Wolfdale looks a bit more attractive for my needs :cool:Quote:
Originally Posted by informal
I don't think a single sample (of each model) is enough to draw conclusions regarding overclockability. When I see 8 Yorkfields out of 10 overclocking badly I will start to draw conclusions.Quote:
Originally Posted by Eastcoasthandle
It does look a bit disconcerting though...
Hot damn! Too bad it's yorkfield vs kentsfield. Wait, 46W less? :)Quote:
Originally Posted by Jacky
For my needs I don't need to draw such a conclusion.Quote:
Originally Posted by Theli
maybe that's why AMD doesn't give away any ES...Quote:
Originally Posted by Theli
NEVER draw too many conclusions about overclockability of a chip just from the ES. If the ES oc well then the chip will probably overclock great
Even though we are on an overclocking board, the main constraint for high speed bins is TDP (temperature) and maybe power draw but not overclockability. (and that is what counts for Intel the most)
Just think about the g0 stepping, it reduced power consumption by almost 10 watts and allows for 200-400mhz higher average overclocks and this is just ONE stepping (one "respin"), but those ES may be early silicon - thus 2-3 "respins" away from final silicon and NOT optimized for overclockability.
The only thing that doesn't change much from early to late silicon is IPC and general chip-architecture, TDP/power draw should go down and OC will go up considerably (thinking about speedpath problems/transistor level optimizations).
edit:
corrected minor mistakes
look, an " early conroe/merom ES OCs bad thread" so let's wait for more results before we make conclusions prematurely
Admittedly I was not around here when the first Conroe ES's were making their rounds, so that's another reason for me to not draw conclusions on overclockability.
I would say though that the temperature difference looks promising.
Wolfdale at over 4 GHz on air with a first run engineering sample and a locked 7X multi seams VERY promising to me.
:rofl: :clap: clever.Quote:
Originally Posted by informal
So, what does AMD have that can do that same sluggish time LOL!?
Most enthusiast know that early boards aren't tweaked or Performance boards weren't even used, firmware is still too RAW and most software is not optimized. Just as the first SP wasn't optimized for Conroe, oh brother. As these things are updated, scores will improved. The same concept applied to AMD and their 2900 on the Video card side applies here as well.
AMD's best hope is that software makers don't optimize for SSE4;)
http://forum.coolaler.com/showpost.p...4&postcount=55
45nm gave me the feeling :
-Ultra-low temperature (hands almost touch radiator lacks feeling so hot ... TAT not support it is not possible to show)
-Lower voltage and low-power! ~ ~ Default 1.15v However 2.33G
We all need to take into consideration another thing..
It's an MSI board!
well, the people have spoken (if you check there are usually only 100 members and some huge amount of guests on) and it looks like almost all of us agree that the benchmarks are legit
looks like an 'easy' transition by intel to 45nm low k gate..hafnium...:)
good chips with low tdp
another small improvement.
im looking forward to the quads - i read that intel may release two versions of porkfield in regular 2.33GHz(333x7) and extreme @ 3.33GHz(333x10)
the extreme part probably being extremely expensive vs affordable 2.33 part.
...but i can imagine some 8x and 9x versions in time - surely it's inevitable...?
if the plethora of current c2d's is anything to go by :lol: