Sign in with
Sign up | Sign in
Your question

3GHZ OPTERONS SPOTTED

Tags:
Last response: in CPUs
Share
March 30, 2006 8:23:49 PM

3GHz single cores; who cares?
Related resources
Can't find your answer ? Ask !
March 30, 2006 8:46:16 PM

Quote:
3GHz single cores; who cares?


Because you can have 8 Single-Core 3Ghz Opterons, that's why.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
March 30, 2006 8:46:54 PM

March 30, 2006 8:48:23 PM

So now that AMD is sort of pushing the Ghz race a little, does this mean they'll ramp up their marketing too?

Intel had their little blue men. I wonder if AMD will use their little green ....

1. Martians
2. Cards
3. Boogers


:lol:  :lol:  :lol:  :lol: 
March 30, 2006 9:04:28 PM

Quote:



LMFAO
March 30, 2006 9:09:38 PM

Quote:
3GHz single cores; who cares?


Because you can have 8 Single-Core 3Ghz Opterons, that's why.

~~Mad Mod Mike, pimpin' the world 1 rig at a timeOr 16 2.8GHz Opterons?
March 30, 2006 9:10:17 PM

Quote:



mmmmmm..i want...


anyways

oo

3Ghz
March 30, 2006 9:10:55 PM

Quote:
3GHz single cores; who cares?


Because you can have 8 Single-Core 3Ghz Opterons, that's why.

~~Mad Mod Mike, pimpin' the world 1 rig at a timeOr 16 2.8GHz Opterons?

Meh, it's more of a breakthrough than a "buy this", it shows AMD can reach 3GHz if necessary and beyond.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
March 30, 2006 9:11:27 PM

What about them ooompla's?

ooompla, ooompla, oompladee dooo, if you are wise, you listen to me..

:oops:  . o O <(NM!! I can't remember the song.)
March 30, 2006 9:25:33 PM

maybe this is my own ignorance but can you enlighten me on something?

its obvious amd can beef up the ghz, so y dont they?

i dont wanna start a fan boy flame on, im being serious.

is there some hardware bottleneck currently? or do they simply not need the extra speed?

obviously the fx-60 is the fastest, but y not make it 3+ghz for the hell of it?
March 30, 2006 9:31:10 PM

AMD's current dual cores have trouble overclocking to 3GHz, much less 10 to 20% over 3GHz, which would be needed for manu to release a stable chip at 3GHz...
I have trouble breaking 2.7GHz with my Opteron 175 stock at 2.2GHz.
March 30, 2006 9:32:44 PM

~~~~~~Attention posters!~~~~~~

Quote:
3GHZ OPTERONS SPOTTED


Turn off the dang caplock!!! Just about every thread's name in this section is capitalized! It doesn't do any good! Turn it off!
March 30, 2006 9:36:21 PM

right, but im curious to what is physically stopping it from goin as high as 3ghz.

is it the density, the type of material they use, etc?

basically what im asking is whats the major physical difference between materials that the processor manufactures use?

i understand y the two excel in different areas such as gaming (amd) and highly intensive business apps (intel)
March 30, 2006 9:37:59 PM

:roll: . o O (okay okay okay, I got a question)

Why don't AMD (or Intel) just make one Processor, and you decide what clock rate to set it?

I don't want marketing prices for answers... :p 
March 30, 2006 9:50:14 PM

that could go either way.

why dont they jsut give you the parts and you put a processor together yourself? same thing.

kids/people with no knowledge would be frying em up like no other. which would cause people to buy another processor (good for manufacturer) or not buy it at all (bad for manufacturer).

they will never give up their pricing ladder! never! its pure profit.

thats like asking hershy's to make a plain candy bar so you can add your own toppings, when they can simply charge you more for a few peanuts, litterally, wtf.
a b à CPUs
March 30, 2006 9:53:44 PM

3ghz, i doubt it can go much further on 90nm, and heat is another thing

woodcrest comes out at 2.93ghz and would eat it 4way minimum, remember DIB is there for it.

who cares?
March 30, 2006 9:57:18 PM

I think it has to do with the fact that with AMD having the OMC, they simply don't have the headroom to clock much higher. AMDs will overclock but, generally, Intels overclock higher because they have the memory controller on the board instead of the processor.

I also believe it has to do with efficiency. With their current architecture, the higher the clocks, the less efficient they are which applies to any processors in all reality. AMD doesn't want to generate excessive heat and waste as little of energy as possible that of which, they are known to be very efficient.

I'm sure someone will elaborate more on this as, i'm usually on the Intel side of things but, i'm using AMDs now as, they have better architecture.
March 30, 2006 9:59:00 PM

Heheh... give me the parts to put together a micro processor.

I don't think there would be any companies, if that was the case, as long as I had a cheap supply of materials. I'd have my own fabrication faculity. I would have the mind of an engineer, and would know the limits of it.

It would be a hell of allot cheaper to do it that way, I could burn up more processors then machines that I built in the past, and not spend anywhere near to what I've spent up today, in just keeping up.

:oops:  But I don' t think I'm that smart. And I doubt MS would talk to me.

:cry: 
March 30, 2006 10:02:25 PM

thanks alot luminaris that was very enlightening. i had speculated about the memory controller being on board but was waiting to hear from someone else.

it makes sense to a point that adding more parts to the core of a cpu could possibly hinder it if not as much as aid it.
March 30, 2006 10:04:28 PM

Quote:
Heheh... give me the parts to put together a micro processor.

I don't think there would be any companies, if that was the case, as long as I had a cheap supply of materials. I'd have my own fabrication faculity. I would have the mind of an engineer, and would know the limits of it.

It would be a hell of allot cheaper to do it that way, I could burn up more processors then machines that I built in the past, and not spend anywhere near to what I've spent up today, in just keeping up.

:oops:  But I don' t think I'm that smart. And I doubt MS would talk to me.

:cry: 


haha. and to add the idea of an unset clock on a cpu is cool to me, i didnt mean to sound as negative as my post did :D 
March 31, 2006 1:20:30 AM

Quote:
AMD's current dual cores have trouble overclocking to 3GHz, much less 10 to 20% over 3GHz, which would be needed for manu to release a stable chip at 3GHz...
I have trouble breaking 2.7GHz with my Opteron 175 stock at 2.2GHz.


This is for the 2 idiots RichPLS and Apache_Lives:

Overall: AMD FX-57 MemoryExtremeTeam - 4101.98 MHz http://valid.x86-secret.com/show_oc?id=56028

1.4 GHz (Applebred): cpulloverclock - 3031.14MHz http://valid.x86-secret.com/show_oc.php?id=40070
1.8 GHz (Applebred): fr@me&Xupyp1 - 2881 Mhz http://valid.x86-secret.com/show_oc?id=73482
1700+ (Toro): Misteroadster - 3227 MHz http://valid.x86-secret.com/show_oc?id=44359
1800+ (Toro): boblemagnifique - 3234.8 MHz http://www.membres.lycos.fr/boblemag...00/3235MHz.png
2500+ (Barton): NewBeetle - 3578.4MHz - http://resources.vr-zone.com/ocdb/da...tel_or_amd=amd
2600+ (Barton): pietro - 3466.3MHz - http://resources.vr-zone.com/ocdb/da...tel_or_amd=amd
2800+ (Clawhammer): setyotomo & ilyas - 3204.14 MHz http://valid.x86-secret.com/show_oc.php?id=18206
3000+ (Barton): Memesama - 3613.96MHz - http://resources.vr-zone.com/ocdb/vi...=pjpeg&which=i
3000+ (Venice 754): amazetestlab - 3090.28MHz - http://valid.x86-secret.com/show_oc.php?id=65486
3000+ (Oakville): firefly - 2700.13MHz - http://valid.x86-secret.com/show_oc.php?id=65488
3000+ (NewCastle): Ssilencer - 3276.60 MHz http://img391.imageshack.us/img391/4638/32765gm.gif
3000+ (ClawHammer): fr@me - 3156 MHz http://valid.x86-secret.com/show_oc?id=65931
3000+ (Winchester): easypanic - 3412.61 MHz http://valid.x86-secret.com/show_oc?id=15933
3000+ (Venice): cpulloverclock - 3644.18 MHz http://valid.x86-secret.com/show_oc?id=41857
3200+ (NewCastle): .::Maddog::. & mosin_nagant - 3024MHz http://valid.x86-secret.com/show_oc.php?id=22563
3200+ (Manchester): ryba - 3405.67 MHz http://valid.x86-secret.com/show_oc?id=58776
3200+ (Winchester): NewBeetle - 3463 MHz http://cal930.sakura.ne.jp/cgi-bin/c...le/3463MHz.gif
3200+ (Venice): s7e9h3n - 3723.6MHz - http://www.xtremesystems.org/forums/...8&postcount=12
3400+ (ClawHammer): eshbach & Evil Spork - 3320.71 MHz http://valid.x86-secret.com/show_oc?id=12416
3400+ (NewCastle): Jrocket - 3359.59 MHz http://valid.x86-secret.com/show_oc?id=18225
3400+ (Venice): damarble - 2850.66 MHz http://valid.x86-secret.com/show_oc.php?id=67088
3500+ (Manchester): Willis - 3205.02 MHz http://valid.x86-secret.com/show_oc?id=79721
3500+ (ClawHammer):o jdr2001 & Silence - 3521.6 MHz http://valid.x86-secret.com/show_oc.php?id=49904
3500+ (NewCastle): |RickY| - 3604.0 MHz http://www.teampuss.com/forums/files/3500caa2c_234.jpg
3500+ (Winchester): Boogotop - 3271,70 MHz http://www.vr-zone.com/ocdb/data.php...tel_or_amd=amd
3500+ (Venice): cpulloverclock - 3571.42 MHz http://valid.x86-secret.com/show_oc?id=46575
3500+ (SanDiego):
3700+ (ClawHammer): twillius_basic - 3331.38 MHz http://valid.x86-secret.com/show_oc.php?id=16530
3700+ (Newark): (Newark): twillius - 3577.35 MHz http://valid.x86-secret.com/show_oc.php?id=25406
3700+ (SanDiego): Misteroadster - 3704.71 MHz http://valid.x86-secret.com/show_oc?id=38280
3800+ (Venice): PC Ice - 3878.05 MHz http://valid.x86-secret.com/show_oc.php?id=10699
3800+ (Toledo): destro404 - 3104.89MHz - http://valid.x86-secret.com/show_oc.php?id=72168
3800+ (NewCastle): Shamino - 3201MHz http://www.vr-zone.com/ocdb/data.php...tel_or_amd=amd
4000+ (Newark): metalman2785 - 3331. 1MHz http://www.xtremesystems.org/forums/...&postcount=702
4000+ (SanDiego): El Snorro - 3750.83 MHz http://valid.x86-secret.com/show_oc?id=76885
4000+ (ClawHammer): gubben - 3614.77 MHz http://valid.x86-secret.com/show_oc?id=25221
FX-53 (ClawHammer): |RickY| & Pedro Rocha - 3602.8 MHz http://resources.vr-zone.com/ocdb/da...tel_or_amd=amd
FX-55 (San Diego): Shamino - 4026.33 MHz http://valid.x86-secret.com/show_oc.php?id=24489
FX-55 (ClawHammer): |RickY|&Pedro Rocha - 4005.23 MHz http://valid.x86-secret.com/show_oc.php?id=10680
FX-57: MemoryExtremeTeam - 4101.98 MHz http://valid.x86-secret.com/show_oc?id=56028
FX-60: mikeguava - 3868.82MHz - http://valid.x86-secret.com/show_oc.php?id=75052
X2 3800+: Hoschi - 3602.6MHz http://valid.x86-secret.com/show_oc?id=62810
X2 4200+: sson74 - 3557.12 MHz http://valid.x86-secret.com/show_oc.php?id=80775
X2 4400+: Fosco - 3408.27 MHz http://valid.x86-secret.com/show_oc.php?id=62072
X2 4600+: Onepagebook - 3687.15 MHz http://valid.x86-secret.com/show_oc.php?id=78999
X2 4800+: MikeGuava - 3560MHz http://valid.x86-secret.com/show_oc.php?id=72471
Opteron 144 (Venus): Overcrash - 3657.8 MHz http://valid.x86-secret.com/show_oc?id=75334
Opteron 146 (Venus): MOSO - 3813 Mhz - http://valid.x86-secret.com/show_oc.php?id=70263
Opteron 148 (Venus): Daniel - 3787.85 MHz http://valid.x86-secret.com/show_oc.php?id=47280
Opteron 150 (Venus): Predator - 3600.9MHz - http://img25.imageshack.us/img25/799...0m373006ax.jpg
Opteron 154 (Venus): s7e9h3n - 4005.19MHz - http://valid.x86-secret.com/show_oc?id=70094
Opteron 165 (Toledo): Nephilim - 3454.8 MHz http://valid.x86-secret.com/show_oc?id=78027
Opteron 170 (Toledo): Absolute_0 - 3473.7 Mhz http://valid.x86-secret.com/show_oc?id=74927
Opteron 175 (Toledo): ixtapalapaquetl - 3420.73 MHz http://valid.x86-secret.com/show_oc?id=80573
Opteron 180 (Toledo): cupra - 3450.99 MHz http://valid.x86-secret.com/show_oc?id=76888
Turion MT-30 (Lancaster: Badong - 3000 MHz http://valid.x86-secret.com/show_oc.php?id=81677
Turion MT-40 (Lancaster): ryan.crest - 2935.4 MHz http://www.xtremesystems.org/forums/...1&d=1141651189
Sempron 2800+ (Paris): turez - 2498.25 MHz http://valid.x86-secret.com/show_oc.php?id=38643
Sempron Mobile 3000+ : hs - 2972.41 Mhz http://valid.x86-secret.com/show_oc?id=69610
Sempron64 2500+ (Palermo): JunQiang - 3003.86 MHz http://valid.x86-secret.com/show_oc.php?id=75385
Sempron64 2600+ (Palermo): pawell262 - 3220.99 MHz http://valid.x86-secret.com/show_oc.php?id=67685
Sempron64 2800+ (Palermo): fr@me - 3287.69 MHz http://valid.x86-secret.com/show_oc?id=55087
Sempron64 3000+ (Palermo): fr@me - 3133.20 MHz http://overclockers.ru/cpubase/?action=details&id=12402
Sempron64 3000+ (Winchester): TiN_EOF - 3031.20 Mhz http://www.topmods.net/tin/3031.png
Sempron64 3000+ (Palermo): pr1nce - 3208.18 MHz http://valid.x86-secret.com/show_oc?id=75376
Sempron64 3100+ (Palermo): STEELBALLZ/MARTAKIS - 3144.31 MHz http://valid.x86-secret.com/show_oc.php?id=75165
Sempron64 3200+: AeGis - 3065.32 MHz http://valid.x86-secret.com/show_oc.php?id=78404
Sempron64 3300+ (Palermo): XFXGeforced - 2843.9 MHz http://valid.x86-secret.com/show_oc.php?id=78287
Sempron64 3400+: fr@me - 2928.95 MHz http://valid.x86-secret.com/show_oc?id=66864

I can break 2.7 on a Opty 165, if you can't do that on a 175, you're a newbie...plain and simple, and you need to stop talking.

BTW: DIB doesn't improve performance Apache, having 2 FSB's to 1 Northbridge, yea that works :roll:.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
March 31, 2006 3:58:12 AM

Actually it has to do more with stages and pipelines.

Pentium 4's netburst has an insane number of stages which allows for higher clock speeds. As we've discussed a billion times before, AMD's focus is IPC. One way a high IPC is achieved is through shorter pipelines and fewer stages. Intel's focus with the Pentium 4 was primarily clock speed and large L2 caches. Every successive Pentium 4 had more and more stages which allowed for higher clock speeds. Of course that's changed w/ the Pentium-M, Core Duo, and Conroe.

No one design is really superior than the other IMO. It takes an engineering feet to design IC's that are capable of high clock speeds. But there are certainly major limitations to just ramping up clock speeds, current leakage being one of them.

-mpjesse
March 31, 2006 4:11:19 AM

Quote:
AMD's current dual cores have trouble overclocking to 3GHz, much less 10 to 20% over 3GHz, which would be needed for manu to release a stable chip at 3GHz...
I have trouble breaking 2.7GHz with my Opteron 175 stock at 2.2GHz.


This is for the 2 idiots RichPLS and Apache_Lives:

Overall: AMD FX-57 MemoryExtremeTeam - 4101.98 MHz http://valid.x86-secret.com/show_oc?id=56028

1.4 GHz (Applebred): cpulloverclock - 3031.14MHz http://valid.x86-secret.com/show_oc.php?id=40070
1.8 GHz (Applebred): fr@me&Xupyp1 - 2881 Mhz http://valid.x86-secret.com/show_oc?id=73482
1700+ (Toro): Misteroadster - 3227 MHz http://valid.x86-secret.com/show_oc?id=44359
1800+ (Toro): boblemagnifique - 3234.8 MHz http://www.membres.lycos.fr/boblemag...00/3235MHz.png
2500+ (Barton): NewBeetle - 3578.4MHz - http://resources.vr-zone.com/ocdb/da...tel_or_amd=amd
2600+ (Barton): pietro - 3466.3MHz - http://resources.vr-zone.com/ocdb/da...tel_or_amd=amd
2800+ (Clawhammer): setyotomo & ilyas - 3204.14 MHz http://valid.x86-secret.com/show_oc.php?id=18206
3000+ (Barton): Memesama - 3613.96MHz - http://resources.vr-zone.com/ocdb/vi...=pjpeg&which=i
3000+ (Venice 754): amazetestlab - 3090.28MHz - http://valid.x86-secret.com/show_oc.php?id=65486
3000+ (Oakville): firefly - 2700.13MHz - http://valid.x86-secret.com/show_oc.php?id=65488
3000+ (NewCastle): Ssilencer - 3276.60 MHz http://img391.imageshack.us/img391/4638/32765gm.gif
3000+ (ClawHammer): fr@me - 3156 MHz http://valid.x86-secret.com/show_oc?id=65931
3000+ (Winchester): easypanic - 3412.61 MHz http://valid.x86-secret.com/show_oc?id=15933
3000+ (Venice): cpulloverclock - 3644.18 MHz http://valid.x86-secret.com/show_oc?id=41857
3200+ (NewCastle): .::Maddog::. & mosin_nagant - 3024MHz http://valid.x86-secret.com/show_oc.php?id=22563
3200+ (Manchester): ryba - 3405.67 MHz http://valid.x86-secret.com/show_oc?id=58776
3200+ (Winchester): NewBeetle - 3463 MHz http://cal930.sakura.ne.jp/cgi-bin/c...le/3463MHz.gif
3200+ (Venice): s7e9h3n - 3723.6MHz - http://www.xtremesystems.org/forums/...8&postcount=12
3400+ (ClawHammer): eshbach & Evil Spork - 3320.71 MHz http://valid.x86-secret.com/show_oc?id=12416
3400+ (NewCastle): Jrocket - 3359.59 MHz http://valid.x86-secret.com/show_oc?id=18225
3400+ (Venice): damarble - 2850.66 MHz http://valid.x86-secret.com/show_oc.php?id=67088
3500+ (Manchester): Willis - 3205.02 MHz http://valid.x86-secret.com/show_oc?id=79721
3500+ (ClawHammer):o jdr2001 & Silence - 3521.6 MHz http://valid.x86-secret.com/show_oc.php?id=49904
3500+ (NewCastle): |RickY| - 3604.0 MHz http://www.teampuss.com/forums/files/3500caa2c_234.jpg
3500+ (Winchester): Boogotop - 3271,70 MHz http://www.vr-zone.com/ocdb/data.php...tel_or_amd=amd
3500+ (Venice): cpulloverclock - 3571.42 MHz http://valid.x86-secret.com/show_oc?id=46575
3500+ (SanDiego):
3700+ (ClawHammer): twillius_basic - 3331.38 MHz http://valid.x86-secret.com/show_oc.php?id=16530
3700+ (Newark): (Newark): twillius - 3577.35 MHz http://valid.x86-secret.com/show_oc.php?id=25406
3700+ (SanDiego): Misteroadster - 3704.71 MHz http://valid.x86-secret.com/show_oc?id=38280
3800+ (Venice): PC Ice - 3878.05 MHz http://valid.x86-secret.com/show_oc.php?id=10699
3800+ (Toledo): destro404 - 3104.89MHz - http://valid.x86-secret.com/show_oc.php?id=72168
3800+ (NewCastle): Shamino - 3201MHz http://www.vr-zone.com/ocdb/data.php...tel_or_amd=amd
4000+ (Newark): metalman2785 - 3331. 1MHz http://www.xtremesystems.org/forums/...&postcount=702
4000+ (SanDiego): El Snorro - 3750.83 MHz http://valid.x86-secret.com/show_oc?id=76885
4000+ (ClawHammer): gubben - 3614.77 MHz http://valid.x86-secret.com/show_oc?id=25221
FX-53 (ClawHammer): |RickY| & Pedro Rocha - 3602.8 MHz http://resources.vr-zone.com/ocdb/da...tel_or_amd=amd
FX-55 (San Diego): Shamino - 4026.33 MHz http://valid.x86-secret.com/show_oc.php?id=24489
FX-55 (ClawHammer): |RickY|&Pedro Rocha - 4005.23 MHz http://valid.x86-secret.com/show_oc.php?id=10680
FX-57: MemoryExtremeTeam - 4101.98 MHz http://valid.x86-secret.com/show_oc?id=56028
FX-60: mikeguava - 3868.82MHz - http://valid.x86-secret.com/show_oc.php?id=75052
X2 3800+: Hoschi - 3602.6MHz http://valid.x86-secret.com/show_oc?id=62810
X2 4200+: sson74 - 3557.12 MHz http://valid.x86-secret.com/show_oc.php?id=80775
X2 4400+: Fosco - 3408.27 MHz http://valid.x86-secret.com/show_oc.php?id=62072
X2 4600+: Onepagebook - 3687.15 MHz http://valid.x86-secret.com/show_oc.php?id=78999
X2 4800+: MikeGuava - 3560MHz http://valid.x86-secret.com/show_oc.php?id=72471
Opteron 144 (Venus): Overcrash - 3657.8 MHz http://valid.x86-secret.com/show_oc?id=75334
Opteron 146 (Venus): MOSO - 3813 Mhz - http://valid.x86-secret.com/show_oc.php?id=70263
Opteron 148 (Venus): Daniel - 3787.85 MHz http://valid.x86-secret.com/show_oc.php?id=47280
Opteron 150 (Venus): Predator - 3600.9MHz - http://img25.imageshack.us/img25/799...0m373006ax.jpg
Opteron 154 (Venus): s7e9h3n - 4005.19MHz - http://valid.x86-secret.com/show_oc?id=70094
Opteron 165 (Toledo): Nephilim - 3454.8 MHz http://valid.x86-secret.com/show_oc?id=78027
Opteron 170 (Toledo): Absolute_0 - 3473.7 Mhz http://valid.x86-secret.com/show_oc?id=74927
Opteron 175 (Toledo): ixtapalapaquetl - 3420.73 MHz http://valid.x86-secret.com/show_oc?id=80573
Opteron 180 (Toledo): cupra - 3450.99 MHz http://valid.x86-secret.com/show_oc?id=76888
Turion MT-30 (Lancaster: Badong - 3000 MHz http://valid.x86-secret.com/show_oc.php?id=81677
Turion MT-40 (Lancaster): ryan.crest - 2935.4 MHz http://www.xtremesystems.org/forums/...1&d=1141651189
Sempron 2800+ (Paris): turez - 2498.25 MHz http://valid.x86-secret.com/show_oc.php?id=38643
Sempron Mobile 3000+ : hs - 2972.41 Mhz http://valid.x86-secret.com/show_oc?id=69610
Sempron64 2500+ (Palermo): JunQiang - 3003.86 MHz http://valid.x86-secret.com/show_oc.php?id=75385
Sempron64 2600+ (Palermo): pawell262 - 3220.99 MHz http://valid.x86-secret.com/show_oc.php?id=67685
Sempron64 2800+ (Palermo): fr@me - 3287.69 MHz http://valid.x86-secret.com/show_oc?id=55087
Sempron64 3000+ (Palermo): fr@me - 3133.20 MHz http://overclockers.ru/cpubase/?action=details&id=12402
Sempron64 3000+ (Winchester): TiN_EOF - 3031.20 Mhz http://www.topmods.net/tin/3031.png
Sempron64 3000+ (Palermo): pr1nce - 3208.18 MHz http://valid.x86-secret.com/show_oc?id=75376
Sempron64 3100+ (Palermo): STEELBALLZ/MARTAKIS - 3144.31 MHz http://valid.x86-secret.com/show_oc.php?id=75165
Sempron64 3200+: AeGis - 3065.32 MHz http://valid.x86-secret.com/show_oc.php?id=78404
Sempron64 3300+ (Palermo): XFXGeforced - 2843.9 MHz http://valid.x86-secret.com/show_oc.php?id=78287
Sempron64 3400+: fr@me - 2928.95 MHz http://valid.x86-secret.com/show_oc?id=66864

I can break 2.7 on a Opty 165, if you can't do that on a 175, you're a newbie...plain and simple, and you need to stop talking.

BTW: DIB doesn't improve performance Apache, having 2 FSB's to 1 Northbridge, yea that works :roll:.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Damn; those are some high overclocks :) . I'd be happy with my P4 at 4GHz; i can't do that because of this shitty proprietary motherboard, but if I get a real board, watch out. But even a 4GHz, an FX57 at stock speeds would pwn mine. Only a 7GHz P4 could perform better than that 4GHz FX57.
March 31, 2006 4:14:03 AM

No Pentium 4 has reached passed 7GHz and been stable in Windows, the FX-57 is, so currently, AMD has the fastest STABLE overclocks, stick that in your pipe Fanboys and smoke it.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
March 31, 2006 4:46:18 AM

Quote:
Not so fast man ... Intels do overclock past 7Ghz and they can be very stable.

http://valid.x86-secret.com/records.php


So there's a SS in CPU-Z, even that kid over in the overclocking forum got 6.2GHz and a CPU-Z shot (if that's real), running CPU-Z is anything but breathtaking.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
March 31, 2006 6:14:45 AM

Quote:
3GHz single cores; who cares?


Because you can have 8 Single-Core 3Ghz Opterons, that's why.

~~Mad Mod Mike, pimpin' the world 1 rig at a timeOr 16 2.8GHz Opterons?WTF you would really waste your money on 16 processors?
March 31, 2006 6:18:09 AM

WOW, this is definitly for rich morons...$1514 for singlecore Opteron....
no thanks.

Mike, about stability:
The SuperPi1M WR goes to Prescott:

SuperPi 17.797s @ 7456MHz
Max OC @ 7638MHz

http://www.xtremesystems.org/forums/showpost.php?p=1208...
K8 are good, but 7.5Ghz are damn 7.5GHz and that is fast and the best.

Yesterday Kingpin made a the best SuperPi1M score for AMD 20.891. He used Athlon64 FX-57 at 4044MHz, I wonder at what temp this good chip has its coldbug. Anyway, its score is far away from 17.797...

http://www.xtremesystems.org/forums/showthread.php?t=94...
March 31, 2006 7:21:47 AM

I suspect it is the VRM, etc on the mainboards.

They only spec'd it so high.

That and so few AMD64 processors on 90nm can reach 3 GHz with 'normal' cooling.

They want gamers buying the Athlon 64FX series anyway.

Quote:
WTF you would really waste your money on 16 processors?


It is 8 x dual-core Opterons.

http://www.vmware.com - just an example.

It is designed for business purposes, and often 'saves' money.
March 31, 2006 8:14:37 AM

Quote:
3GHz single cores; who cares?


Because you can have 8 Single-Core 3Ghz Opterons, that's why.

~~Mad Mod Mike, pimpin' the world 1 rig at a timeOr 16 2.8GHz Opterons?WTF you would really waste your money on 16 processors?

I wish I had that much money...

And a processor at $1500+ is really not going to affect anyone for some time. Excepting the stupid rich people who can afford things like this that I can't... :evil: 
a c 99 à CPUs
March 31, 2006 12:42:05 PM

That's pretty neat as it shows the maximum capabilities of the chips. What I would like to see is a chart like that with the maximum overclocks on the stock HSF, with the full-load temp at least 5 C below shutdown temp at all times, and with the same stability as at stock. That would be a great indicator of what somebody could reasonably extract from an otherwise stock chip and be able to use it like a stock chip. It would also be a decent indicator of how much more headroom AMD has on the current process as AMD isn't going to be shipping chips that require water cooling to keep below 50 C at idle.
March 31, 2006 12:51:20 PM

On stock HS, AMD dual core has mucho trouble breaking 3GHz... bottom line.
Intel dual core on stock HS easily 4+GHz
March 31, 2006 1:56:40 PM

Quote:


Mike, about stability:
The SuperPi1M WR goes to Prescott:

SuperPi 17.797s @ 7456MHz
Max OC @ 7638MHz


Makes you wonder about the record-breaking power consumption to achieve that marvellous feat you're talking about.
March 31, 2006 5:31:40 PM

HOT!
March 31, 2006 5:33:52 PM

Quote:
Actually it has to do more with stages and pipelines.

Pentium 4's netburst has an insane number of stages which allows for higher clock speeds. As we've discussed a billion times before, AMD's focus is IPC. One way a high IPC is achieved is through shorter pipelines and fewer stages. Intel's focus with the Pentium 4 was primarily clock speed and large L2 caches. Every successive Pentium 4 had more and more stages which allowed for higher clock speeds. Of course that's changed w/ the Pentium-M, Core Duo, and Conroe.

No one design is really superior than the other IMO. It takes an engineering feet to design IC's that are capable of high clock speeds. But there are certainly major limitations to just ramping up clock speeds, current leakage being one of them.

-mpjesse


I believe that's the answer, although Intel's Core Architecture is actually a new architecture and - although much closer to AMD's performance-per-Watt approach - it's an overall departure, no doubt superior to the current AMD's K8. [Competitive] technological progression so demands it (as everybody expects K8L to be superior to K8...).

Do you think the IMC can also be a limitation to the frequency scale up?


Cheers!
March 31, 2006 6:51:01 PM

Quote:
Does anyone have any more details on K8L? It is weird how this is getting banted around, then the announcement on Tom's that it was the core going into the quad core release in Q1 07.


Hello!

Nope. But I've been around other forums, as well. I'll let you know if I find something relevant.

(I've also been busy trying to understand my own idea [p-IMC]... not being a specialist, it's been haaard! :D  )


Cheers!
March 31, 2006 6:57:09 PM

Quote:
3GHz single cores; who cares?


Because you can have 8 Single-Core 3Ghz Opterons, that's why.

~~Mad Mod Mike, pimpin' the world 1 rig at a timeOr 16 2.8GHz Opterons?WTF you would really waste your money on 16 processors?

You think gamers are the only people who need computers? Got a database server that would be run into the ground if it had ONLY 16 cpus.
March 31, 2006 7:33:13 PM

Merom, Conroe, Woodcrest, etc will have +80% the IPC of NetBurst.

eg: 5.42 IPC vs 3.54 (including HyperThreading, only 3 IPC w/o) IPC

Notes:
- This code has lots of loops (so do game engines, but they don't fit within the cache of the processor :p )
- 5.42 IPC is my relative prediction for Conroe, (these are all per core btw)
- 3.54 IPC is a 'best case' from NetBurst really, it is often more like 3 IPC.
- I personally put K8 at apx 4.93 IPC for these comparisons.
- This is depending on the code, and thus instructions, used, etc, I've factored in L1 caches but only slightly, here btw. The figures could be far lower depending on the code, but the important thing is that they are relative figures. 8)

Even a ~ 2 GHz Conroe is going to kick sweet ass at around ~ 21,680 MIPS (+/- 10%). That score includes both cores btw.

The 2.67 GHz model will thus be near 28,910 MIPS, which is 70% the performance of my machine, from Intel, using only two cores (vs four), and thus only requring two (vs four for my rig) isolated threads per running process for maximum performance.

If every processor was just given a MIPS / MFLOPS rating, with the number of cores listed, it would make life much easier, esp once caches, given code, etc are factored in.

So yes, a 3 GHz dual-core AMD64 (be it Athlon 64 or Opteron) will 'just' beat the 2.67 GHz conroe performance wise.
However they'll cost significantly more to build, unless AMD move to a 65nm SOI manufacturing process ASAP. (Which is what they are doing, but they'll have costs to cover).

Although, Conroe clock could be raised +25% to get to 100w TDP at 65nm for Intel, and thus sold for more, aswell as requiring 3.75 GHz (apx) from AMD64 K8 microarchitecture 'to have an equal' from AMD.

This is why I like Merom, Conroe & Woodcrest (and Kentsville) - the designs work at 65w TDP, or less, and also at 112w TDP. It could even do 3.5 GHz at around 112w TDP initially (by my numbers, which are usually highly accurate), this would likely translate into 3.00 GHz and 3.33 GHz parts as they (Intel) want to put out less heat, and thus generate a more 'silent' PC, while using less energy overall. (The world has 4 billion PCs or so, how much power would they all be using, even when idle ?)

Surface area to contact ratio, important for cooling, is on Intels side with these designs, the market also has benefited from PreScott as those 'advanced heatsink' designs can now be ported to Conroe, etc to provide it with far more cooling than it requires.

Imagine only requiring passive cooling for such a CPU at 1.4 - 1.8 GHz, would be excellent for HTPC devices and the like. My instinct tells me Intel want to get into this HTPC & HDTV devices market in the 5 years from 2007 to 2012.
March 31, 2006 8:37:18 PM

Quote:
- I personally put K8 at apx 4.93 IPC for these comparisons.

I put K8 at apx 4.50 IPC.
Quote:
So yes, a 3 GHz dual-core AMD64 (be it Athlon 64 or Opteron) will 'just' beat the 2.67 GHz conroe performance wise.

I don't agree with you. A 3GHz dualcore AMD64 will not beat 2.67 Conroe. It was benchmarked with 2.8GHz dualcore oc-ed FX-60 and was apx 20% better in all test. A 3GHz dc K8 will perform 7.1% faster than 2.8GHz(3*100/2.8=107.1%). Conroe will be 12% faster than 3.0GHz dc K8(120/107,1=112,05%).
Anyway there is no 3.0GHz dualcore Opteron released.

However they'll cost significantly more to build, unless AMD move to a 65nm SOI manufacturing process ASAP. (Which is what they are doing, but they'll have costs to cover).[/quote]
I agree, it is so unAMD like, a $1500+ for a singlecore 3GHz Opteron...
March 31, 2006 9:27:04 PM

Quote:

Anyway there is no 3.0GHz dualcore Opteron released.


There is no Conroe released yet either, AMD may have a few 65nm CPUs out by the time Conroe is released. K8 3.0 GHz dual-cores and Socket AM2 (including new memory controller) are next on their ToDo list, likely out around July - August if estimates are correct. They'll want their 3.2 GHz design, on AM2, to be on par with a 3.0 GHz Conroe, and it will be... assuming no FPU changes on either processor between now and then. (Of course if they both gain +12.5% on FPU it will still have the same relative effect :p )

The Opteron prices (you are only looking at the top dog, try looking down 1 or 2 models btw) will come down by the time Conroe is released they should have 3.2 GHz models and more stock on the market.

FPS != IPC (ie: As the FPU is used in games, Conroe FPU performance has only jumped about about +60% per clock cycle, while IPC has jumped +80% per clock cycle... this may change in the final version of Conroe though, bringing it to +80% on both).

The AMD system they used for testing was handicapped by the ATI chipset performance in Crossfire, where as the Intel chipset used in the Intel system was known for having far better Crossfire scaling than the lower end ATI chipset. If they tested it using just 1 video card on both, without any significant (eg: under 3%) chipset performance differences Conroe would still have won, but not by the same margin as the other tests.

Assuming equal chipset, both dual-core, no FPU improvements to either CPU between now and release (possible), equal scaling in SLI / Crossfire (or simply not using it), etc then a 2.67 GHz Conroe would have performed 'on par' with the 2.80 GHz Athlon 64 FX. The Crossfire scaling (chipset differences) gained them more than +10%. This why they didn't just use one card on each machine, or use SLI (which scales fairly equally regardless of chipset, so long as it is configured correctly on both machines).

Yes, Crossfire was used in both machines.
However, The Crossfire configurations did not scale equally (not even within 3%) between the two machines they demonstrated at IDF. (to Anandtech and others, Anandtech has picked up on this but I don't think they've made any official comments about it yet).

The IPC figures I used are all relative anyway. 8)

So if you put K8 at 4.50 IPC, that would put Conroe at 4.94 IPC, NetBurst at 2.73 IPC, and NetBurst with HyperThreading at 3.23 IPC. (all apx).

The numbers I used where using x64 code btw, using small loops, under an x64 OS.

K8 IPC rises more, as a percentage, from x86 to x64, comparied to NetBurst going from x86 to x64. Conroe is likely to be similar to K8 in this respect.

eg: Going from x86 to x64 the K8 IPC may rise from 4.5 to 4.93, or from 4.11 to 4.5 :p  (relative).

The Intel Pentium D/X 840 (Extreme Edition 840 with HT) gains +20% IPC going from x86 to x64. While a single core Pentium 4 without HT only gains about +4% IPC during the same transition. Oddly HyperThreading benefits to IPC are amplified in a pure x64 environment. (That is x64 OS running x64 software - Not 32-bit software under an x64 OS though).

Note: x64 covers: x86_64, AMD64, EM64T

My predictions thus put K8 / K8L on Socket AM2 at 3.0 GHz performing 'on par' with the 2.80 GHz Conroe in games. Assuming same number of cores, not that game software benefits from 4-way systems anyway, aswell as same video subsystem setup, including using 'on par' chipsets. :p 

To break even with a 2.80 GHz Conroe AMD only need +2.5% performance from 3.0 GHz dual-cores on Socket AM2, which is easily within their reach due to the new memory controller.

This all assumes no major changes (K8L isn't major, maybe more cores though) are made to to either processor. I suspect at least one, if not both, of them is/are going to make improvements to the FPU.

FPU performance not scaling with IPC would be classed as a 'major concern' by the kind of engineers working on these processors IMHO. It would affect performance, and thus profits, so the managers would be notified and 'last prototyping' changes are likely to affect FPU performance. (See top comment in blue, 1st large paragraph above).
March 31, 2006 10:00:04 PM

Quote:
WTF you would really waste your money on 16 processors?
I'm guessing you've never heard of a "Server".
March 31, 2006 10:41:50 PM

Technically, Windows 2003 Enterprise supports up to 64 processors and up to 127GB of RAM...and there are servers that utilise this. Yup, 16 is nothing man. :p  :mrgreen:
March 31, 2006 11:01:33 PM

Quote:
Merom, Conroe, Woodcrest, etc will have +80% the IPC of NetBurst.

eg: 5.42 IPC vs 3.54 (including HyperThreading, only 3 IPC w/o) IPC

Notes:
- This code has lots of loops (so do game engines, but they don't fit within the cache of the processor :p )
- 5.42 IPC is my relative prediction for Conroe, (these are all per core btw)
- 3.54 IPC is a 'best case' from NetBurst really, it is often more like 3 IPC.
- I personally put K8 at apx 4.93 IPC for these comparisons.
- This is depending on the code, and thus instructions, used, etc, I've factored in L1 caches but only slightly, here btw. The figures could be far lower depending on the code, but the important thing is that they are relative figures. 8)


That's interesting. Could you be more specific? i.e., how did you arrive to these numbers?


Cheers!
March 31, 2006 11:34:19 PM

Quote:

So yes, a 3 GHz dual-core AMD64 (be it Athlon 64 or Opteron) will 'just' beat the 2.67 GHz conroe performance wise.
However they'll cost significantly more to build, unless AMD move to a 65nm SOI manufacturing process ASAP. (Which is what they are doing, but they'll have costs to cover).


And, are you certain a move [AMD's] into the 65nm node will bring a ~15% increase in frequency? (All taken into account, that is.).

Quote:
Surface area to contact ratio, important for cooling, is on Intels side with these designs, the market also has benefited from PreScott as those 'advanced heatsink' designs can now be ported to Conroe, etc to provide it with far more cooling than it requires.


I guess this is a point which has barely been addressed but, nevertheless, a more & more relevant one, even if TDP envelopes are becoming more tight. Do you know if there'll be any change in the Conroe's heat spreader, namely on the materials used? (I'm aware this is a precocious question).


Cheers!
April 1, 2006 3:22:43 AM

Quote:
:roll: . o O (okay okay okay, I got a question)

Why don't AMD (or Intel) just make one Processor, and you decide what clock rate to set it?

I don't want marketing prices for answers... :p 


Hey man it is called overclocking if you didnt know. And I would imagine that they don't do that because then they would lose money, and losing money is NOT a good thing if you didn't happen to know. Just generally a bad idea.
April 2, 2006 3:12:23 AM

Quote:
:roll: . o O (okay okay okay, I got a question)

Why don't AMD (or Intel) just make one Processor, and you decide what clock rate to set it?

I don't want marketing prices for answers... :p 


Hey man it is called overclocking if you didnt know. And I would imagine that they don't do that because then they would lose money, and losing money is NOT a good thing if you didn't happen to know. Just generally a bad idea.

Oh.. I see you need to explain something I already know, if you entirely understand my question. :roll:

To explain - rather then buying a product (that you can overclock) why not buy a processor that you can clock yourself, based on your own needs. Surely, any OC'er would push it as far as it could go.

The 'marketing' answer, is basically the only correct answer, which you just re-enstated, since both AMD/Intel milks the cow as much as they can, which does help the stock market.

Quote:
Just generally a bad idea


As far as bad idea, yes, for companies that sell. No, since anyone who OC's will push it as far as it will go, and that there would be only 1 price, rather then ranges of prices on CPUs. :p 

Edit: And it was only a question, not a proposal/idea.
April 2, 2006 6:26:00 AM

Quote:
3GHz single cores; who cares?


Because you can have more than one Opteron in a machine.
!