Friday, September 07, 2007

K10 performance estimates

INQ had some rough numbers. On SPECint_rate_2006, a 2.3GHZ K10 is 21% faster than a 2.66GHZ Clovertown. On SPECfp_rate_2006, the 2.3GHZ K10 is 50% faster than a 2.66GHZ Clovertown.

These numbers translates to 40% IPC advantage in integer and 73% IPC advantage in FP performance.

I expect at 4P level, the 2GHZ K10 should be close to Intel's 2.93GHZ quad in integer performance and will be 50% faster than the Intel 2.93GHZ quad in floating point.

16 Comments:

Blogger Al-Saqer said...

So far, only promises. I am waiting for the cow.

8:10 PM, September 07, 2007  
Blogger Giant said...

AMD is finished. Barcelona is fragged by 1.86Ghz Intel Quad Core:

http://blogs.zdnet.com/Ou/images/4p-intel-vs-amd-server.png

AMD is finished. BK in Q2'08.

9:59 PM, September 07, 2007  
Blogger pointer said...

normally i don't comment base on what the host said as i found no value commentting. Anyway, from the same Inq link, look at the bottom:

* Correction The listed SPECint_rate2006 and SPECfp_rate2006 should be compared to a simulated 2.6GHz Barcelona, not a 2.3GHz.


1st) this is a july article
2nd) simulated results ... :)
3rd) latest leak result in George Ou's page link by Giant
4th) 2 days later, AMD will anyway publish those results
5th) Most other benchmarking sites will publish their own testing result
6th) Those benchmarking sites WILL be labelled by some people here as Intel pomper site if those result are not in AMD's favor ... :)

10:15 PM, September 07, 2007  
Blogger Giant said...

Sharikou isn't even trying anymore. He already posted these old simulated results a long time ago.

AMD BK Q2'08.

10:16 PM, September 07, 2007  
Blogger z said...

giant:

2.0GHz Opti costs $1004, scores 160.
1.86GHz Xeon costs $2301, scores 166.

Would you spend an extra $1297 for +3.75% in your SPECint score ? Use your FUCKING brain.

Oh wait, Intel fanbois have no brain :)

10:23 PM, September 07, 2007  
Blogger anti-Intel guy said...

@z >>> LOL

@giant: I spoke with Theo from L'Inq (again): "[...] Intel needs Nehalem today [...]"

So, mr. Giant, stop posting your BK bullshit on this blog, because you're not funny anymore.

10:40 PM, September 07, 2007  
Blogger Sharikou, Ph. D said...

I talked to Theo from L'Inq just a little while ago.
He said AMD is fucked.

11:06 PM, September 07, 2007  
Blogger Giant said...

This comment has been removed by the author.

11:27 PM, September 07, 2007  
Blogger Giant said...

2.0GHz Opti costs $1004, scores 160.
1.86GHz Xeon costs $2301, scores 166.
Would you spend an extra $1297 for +3.75% in your SPECint score ? Use your FUCKING brain.



Hahaha. I didn't realise the AMD fanboys are such idiots. Why are you comparing a 50W LV Xeon to the fastest Barcelona you can buy? Who's the fucking idiot now?

The competitor to AMD's fastest Barcelona CPU is the $1172 2.13Ghz Tigerton CPU. It's faster than the LV Xeon and about half the price you moron.

Look at AMD's pricing. $1000 for the fastest Barcelona CPU? ahahaha. AMD's ASP is going to drop even more and utterly collapse. Apparently losing $600m a quarter isn't quite enough for AMD! Yes, remember AMD is losing $7M EVERY DAY!

So, anti-intel guy, stop posting your Theo 30K 3Dmnark06 reverse hyperthreading bullshit on this blog, because you're not funny anymore.

AMD BK Q2'08.

11:31 PM, September 07, 2007  
Blogger Hector de J. Ruiz, Ph.D said...

To whom it may concern,

I'm here to tell you that AMD is fucked. Quad Core Opteron is slower than our competition and the fastest Quad Core Opteron CPU we will have by the end of the year is a paltry 2.3Ghz.

But I don't care about that at all. I made a cool 16 million last year. When AMD goes under I'll be enjoying my retirement at my mansion.

I just thought our most loyal supporters like Sharikou, oneexpert and Penix deserved to know the truth.

Regards,

Hector de J. Ruiz, Ph.D

1:30 AM, September 08, 2007  
Blogger anti-Intel guy said...

intel fanbois = a bunch of idiots

1:39 AM, September 08, 2007  
Blogger 13ringinheat said...

Well there you have it.......from the horse's mouth itself. Thanks Hector. :P

Hahhahaha dumbass Z and anti intel guy cant even make simple comparisons now........just another day for delusional AMD fanbois here at sharikou's

1:39 AM, September 08, 2007  
Blogger 13ringinheat said...

Anti intel guy says: intel fanbois = a bunch of idiots

After getting completely owned by giant this is his comeback way to go junior...let me know when you graduate middle school......

Your insults are as pathetic as the company you idolize

1:41 AM, September 08, 2007  
Blogger anti-Intel guy said...

Let's see:
- AMD takes x86 to 64-bits in 2003, Intel follows years later
- AMD brings Direct Connect to market in 2003 and blows out x86 server bandwidth and scalability, Intel follows in 2008/2009
- AMD brings performance-per-watt to the datacenter in 2003, Intel follows in 2006
-AMD brings dual-core to x86 in the datacenter in 2005, Intel slaps two single-cores together the next year. Does a real dual-core two years later
- AMD embeds memory controller in 2003 and kills the memory bottleneck, Intel follows in 2008/2009
-AMD brings quad-core design to the x86 datacenter in 2007, Intel follows in 2008/2009.

As I said before: intel fanbois = a bunch of idiots

1:51 AM, September 08, 2007  
Blogger Giant said...

- AMD takes x86 to 64-bits in 2003, Intel follows years later
- AMD brings Direct Connect to market in 2003 and blows out x86 server bandwidth and scalability, Intel follows in 2008/2009
- AMD brings performance-per-watt to the datacenter in 2003, Intel follows in 2006
-AMD brings dual-core to x86 in the datacenter in 2005, Intel slaps two single-cores together the next year. Does a real dual-core two years later
- AMD embeds memory controller in 2003 and kills the memory bottleneck, Intel follows in 2008/2009
-AMD brings quad-core design to the x86 datacenter in 2007, Intel follows in 2008/2009.


You copied this from George Ou's blog. Anyway, this will still be simple.

AMD takes x86 to 64-bits in 2003, Intel follows years later

The first 64bit CPU was MIPS based, it was released in 1991. Still, AMD had the first 64bit x86 CPU about a year before Intel. They get credit for that.

- AMD brings Direct Connect to market in 2003 and blows out x86 server bandwidth and scalability, Intel follows in 2008/2009

AMD only did this because their own bus couldn't scale past 400mhz when Intel's bus was at 800mhz. Now Intel's FSB is no longer very effective in DP/MP servers. The 1600mhz bus due to be released later this year will be the final upgrade to the FSB. The FSB will be replaced with CSI/QuickPath next year.

AMD brings performance-per-watt to the datacenter in 2003, Intel follows in 2006

Performance per watt is hardly an 'innovation'. Low power CPUs have been around for years. AMD just started talking 'performance per watt' when Netburst hit a thermal wall and the TDP went up to massively high levels. This is more fact that K8 is better than Netburst. I don't know of anyone that would doubt K8's superiority over Netburst.

AMD brings dual-core to x86 in the datacenter in 2005, Intel slaps two single-cores together the next year. Does a real dual-core two years later

Both Intel and AMD released dual core CPUs in 2005. Both were single die solutions. Smithfield was a single die product. Intel didn't start with the dual die approach until 65nm.


- AMD embeds memory controller in 2003 and kills the memory bottleneck, Intel follows in 2008/2009


The 486 had an integrated memory controller back in 1989. This continued until the Pentium processor which actually moved the memory controller onto the northbridge. AMD's 486 CPUs had an integrated memory controller. K5/K6/K7 CPUs did not. Claiming this as some sort of magical new AMD innovation is simply ludicrous.

The first "modern" (post year 2000) CPU to feature an integrated memory controller was the Alpha 21364.

AMD brings quad-core design to the x86 datacenter in 2007, Intel follows in 2008/2009.

This is also totally false. Intel started selling quad core CPUs in November 2006. AMD has not yet released any quad core CPUs. (Unless you count CPUs of the simulated variety :-P)

As I said before: amd fanbois = a bunch of idiots

3:06 AM, September 08, 2007  
Blogger z said...

Why are you comparing a 50W LV Xeon to the fastest Barcelona you can buy?

You compared them: "Barcelona is fragged by 1.86GHz Intel Quad Core".

The competitor to AMD's fastest Barcelona CPU is the $1172 2.13Ghz Tigerton CPU. It's faster than the LV Xeon and about half the price...

...and half the L2 cache, which explains why Tigerton gets trounced by the cheaper 8350 in 4P SPECint_rate2006. Lol! http://blogs.zdnet.com/Ou/?p=735

6:34 PM, September 08, 2007  

Post a Comment

Links to this post:

Create a Link

<< Home