AMD may have 4 instruction decoders up its sleeve.

dorion

Distinguished
Nov 10, 2005
92
0
18,630
http://www.theinquirer.net/?article=32424

Hmm sounds good will it give AMD a fighting chance? K8L still can't fuse macro opps(or is it micro opps?) And they have more thermal output. I just want a nice competitor to Conroe, I don't want an Athlon XP vs. Pentium 4 defeat again.
I want AMD to do good(support the little guy) and Intel to do good(stocks). Wonder what wish I'll get.
 
http://www.theinquirer.net/?article=32424

Hmm sounds good will it give AMD a fighting chance? K8L still can't fuse macro opps(or is it micro opps?) And they have more thermal output. I just want a nice competitor to Conroe, I don't want an Athlon XP vs. Pentium 4 defeat again.
I want AMD to do good(support the little guy) and Intel to do good(stocks). Wonder what wish I'll get.

Maybe not but they are doublign the width of SSE, adding OoO buffers, increasing the fetch to 32KB, adding new 64bit registers and adding siGe to pump clock speed or power efficiency. Intel better hope it oesn't come out.

The design sounds like 40% better perf at dual core (can't compare quad)
 
The design sounds like 40% better perf at dual core (can't compare quad)


Just a second, let me see if I got this straight. If people benchmark Intel machines before release showing a 40% performance increase (over the Pentium D), they can't be right. They're downright wrong. But if AMD even talks about releasing an improved uArch, it's a sudden 40% performance increase. :roll:
 
Well, AMD's policy is to under-promise and over-deliver. Intel is quite the contrary. Remember when they were talking about how they'd have 10 GHz CPUs?
 
AMD Has Been Using 4 Instruction Decoders Since The Introduction Of The AMD K5 Back In 1996, What They Should Due Is Double The The L1 And L2 Caches, Go To .65nm w/SOI, Go To Hypertransport 2.0 (2x32-Bit Lanes Enstead Of 2x16-Bit Lanes), And Design There Own Version Of Memory Disambiguation (I Think Thats How They Spell It) And There Own Version Of Micro-Opps Fusion.

P.S They Do However Plan On Adding L3s To There High End FXs And Opterons
 
The design sounds like 40% better perf at dual core (can't compare quad)


Just a second, let me see if I got this straight. If people benchmark Intel machines before release showing a 40% performance increase (over the Pentium D), they can't be right. They're downright wrong. But if AMD even talks about releasing an improved uArch, it's a sudden 40% performance increase. :roll:


That's not true. When I saw the specs for Conroe I knew it would offer 40% above P4. Maybe not 40% above K8.
 
Intel's Tri-Gate Transistor to boost processor's performance/watt!!!

chart.gif

untitled.jpg

Researchers at Intel have found a better way to insulate circuits, enabling them to save energy as they pack more transistors onto each processor.

Intel could start building chips with these new "tri-gate transistors" by 2010, enabling either a 45 percent increase in speed or a 35 percent reduction in total power used, compared to the company’s current 65-nanometer process transistors, said Mike Mayberry, director of components research and vice president of Intel’s Technology and Manufacturing Group.

The advance could be a powerful sales tool, since power efficiency is a crucial marketing metric for chips in PCs ranging from powerful servers to mobile laptops and handheld PDAs.

Advanced Micro Devices (AMD) has purchased billboard ads that claim its Opteron chip will reduce electricity costs compared to Intel’s Xeon chips. And Sun Microsystems frequently taunts competitors Hewlett-Packard and IBM for using fans to cool their servers, instead of using more efficient chips like Sun’s UltraSPARC T1.

Intel’s new technology would also extend the reach of Moore’s Law, the prediction made 40 years ago by Intel cofounder Gordon Moore that the number of transistors on a chip would double about every two years.

Engineers have recently predicted that trend would soon end, because electricity tends to leak out of tiny wires as chip geometry shrinks below 90 nanometers.

One solution is to build chips with multiple cores that run at slower speeds, since chips leak more electricity and run less efficiently as they run faster than 2GHz. Chip makers from Intel to AMD and Sun have all followed this path.

Another answer could be carbon nanotubes, according to scientists at IBM, who said in March they had built an electronic integrated circuit by combining conventional silicon technology with a carbon nanotube molecule.

But Intel says its tri-gate solution is the best.

"Compared to carbon nanotubes, it is far easier to build," Mayberry said. "The problem with carbon nanotubes is that no one knows how to put them in a particular spot except by moving them one at a time. Even our smallest chips have millions of transistors, so that is an insurmountable challenge."

A tri-gate transistor is a component in the standard complementary metal oxide semiconductor design, but acts as a better "traffic cop" to control the flow of electrons, surrounding each wire on three sides instead of just one.

"It’s better to wrap the gate around, just like it’s easier to block a garden hose by squeezing on all sides than just holding your thumb on one side," he said.

The technology is still on the drawing board, but Intel designers will be able to quickly apply it to new chips someday because they can use existing equipment in the fabrication plants.

"This will be an option for chips somewhere beyond 45 nm—in the 32 or 22 nm mode—so that gives us confidence we can continue scaling Moore’s Law into the next decade," said Mayberry.

Intel has said it will be making more chips with 65-nanometer geometry than with 90 nanometers by the third quarter of 2006, and will move to 45 nanometers in 2007 and 32 nanometers by 2009.



too fast to process a data.. hahahaha... 😵
 
Or just bring back some intelligent conversation. It's been seriously lacking lately. BTW Intel is still trying to get itanium out there to this day :roll:
 
BTW Intel is still trying to get itanium out there to this day

Huh?

Well lets look at their two latest and biggest projects. The Pentium M and Core 2. Neither have overpromised and underdelivered. Just the opposite it would seem.

I'm sry but mixing my quote with someone elses opionion doesnt work for me. What does this have to do with 4 complex decoder's on Rev G and K8L again?
 
What do any of these posts have to do with 4 complex decoder's on Rev G and K8L? Nothing.

Exactly. Now let's end this in peace and let discussion commence. You cant group every AMD fan into this "horde" catogory you've created and I get offended whenever you treat me as such. I will not reply to any further nonsense so say whatever you may to get your last word in AM as I know you'll just be itchin to make me look like a fool one way or another.
 
http://www.theinquirer.net/?article=32424

Hmm sounds good will it give AMD a fighting chance? K8L still can't fuse macro opps(or is it micro opps?) And they have more thermal output. I just want a nice competitor to Conroe, I don't want an Athlon XP vs. Pentium 4 defeat again.
I want AMD to do good(support the little guy) and Intel to do good(stocks). Wonder what wish I'll get.

I believe Rev G with some IPC enhancements increased clockspeeds will put AMD within striking distance of Conroe. I dont think we'll see another major single threaded performance boost until K10 and I'm not entirely sure where Intel is headed other than unifying their socket's with CSI and adding more cores + getting smaller. If nothing else they will put even more cache on a conroe-ish type architecture
 
Wow, didn't see that one coming.

You cant group every AMD fan into this "horde" catogory you've created

I didn't create it and I've done no such thing.

I get offended whenever you treat me as such.

I never have.

I will not reply to any further nonsense so say whatever you may to get your last word in AM as I know you'll just be itchin to make me look like a fool one way or another.

See now thats totally unnecessary and I'm very dissappointed to hear that from you.
 
Wow, didn't see that one coming.

You cant group every AMD fan into this "horde" catogory you've created

I didn't create it and I've done no such thing.

I get offended whenever you treat me as such.

I never have.

I will not reply to any further nonsense so say whatever you may to get your last word in AM as I know you'll just be itchin to make me look like a fool one way or another.

See now thats totally unnecessary and I'm very dissappointed to hear that from you.

Word.
 
Wow, didn't see that one coming.

You cant group every AMD fan into this "horde" catogory you've created

I didn't create it and I've done no such thing.

I get offended whenever you treat me as such.

I never have.

I will not reply to any further nonsense so say whatever you may to get your last word in AM as I know you'll just be itchin to make me look like a fool one way or another.

See now thats totally unnecessary and I'm very dissappointed to hear that from you.

Word.
*sigh* I've always hated being the youngon in the conversation because no matter how hard I try, I cant beat lvl minded people at that game because i'm only starting to learn the mindgames of adulthood and get beat with experience every time.
 
http://www.theinquirer.net/?article=32424

(...) K8L still can't fuse macro opps(or is it micro opps?)

Well, it's both (in Intel's Core uArch); not sure about a floorplan uArch (damn word!); anyway, you could have mentioned the picture the Inq. links to... I've posted it somewhere else and I'll re-post it here (maybe the moods change...):



It's one [supposed] core of the [supposed] K8L Rev H quad-core, seen through Galileo's [supposed] broken last glasses...

And they have more thermal output.

Sorry?!


Cheers!
 
http://www.theinquirer.net/?article=32424

Hmm sounds good will it give AMD a fighting chance? K8L still can't fuse macro opps(or is it micro opps?) And they have more thermal output. I just want a nice competitor to Conroe, I don't want an Athlon XP vs. Pentium 4 defeat again.
I want AMD to do good(support the little guy) and Intel to do good(stocks). Wonder what wish I'll get.

I believe Rev G with some IPC enhancements increased clockspeeds will put AMD within striking distance of Conroe. I dont think we'll see another major single threaded performance boost until K10 and I'm not entirely sure where Intel is headed other than unifying their socket's with CSI and adding more cores + getting smaller. If nothing else they will put even more cache on a conroe-ish type architecture


33% more decoding potential will enable AMD to more than overcome Core 2. Even adding L3 in place of 512 of L2 may increase AM2 by another 5-10%. By the time Intel intros CSI, HTX will be everywhere.

I'm sure AMD has learned enough about x64 to make a horse race of it no matter what Intel does. Next year the new buzzword will be Winx64.
 
Now let's end this in peace and let discussion commence. You cant group every AMD fan into this "horde" catogory you've created and I get offended whenever you treat me as such. I will not reply to any further nonsense so say whatever you may to get your last word in AM as I know you'll just be itchin to make me look like a fool one way or another.

Hey, K8MAN I am dissapointed to read this from you.
I think you are wrong about AM trying to make you a fool. I like his sence of humor, but sometimes it becomes stupid. But he is not offensive to you.
I think that no one included you in the HORDE. The HORDE with its stereotypic ignorant persons is grouping it self alone.
What is wrong with both of you, low rating your posts each to other?