Saturday, September 29, 2007

Intelers' words are disturbing

Spaztic Pizza said: "Oneretard is another example of why some people just shouldn't be allowed to breed.."

I once concluded that Intel's exclusionism has its roots in NAZIsm.

Henry Morgenthau suggested that Germans should be castrated, Roosevelt agreed.

Friday, September 28, 2007

AMD dominates casual gaming performance

You want a basic PC with some 3D gaming capacity so your kids can have some Roller Coaster or reader rabbit fun, or maybe you want to learn to fly with MS Flight Simulator?

With AMD, you don't need to pay $100 for a GPU, AMD's integrated graphics solutions can do it all.

Tuesday, September 25, 2007

K10 available from newegg

Intel won't have this kind of technology until 2009, and that's only on paper. You can get it today from Newegg. Intel engineers often look at K10 with awe: they wonder how AMD can make such a complex CPU work flawlessly.

Be smart, be cool, use AMD K10 technology. Only idiots stick to retarded FSB.

Monday, September 24, 2007

Intel is firing

Not all cylinders, but a bunch of people who hoped to retire there.
FACE Intel adds some fresh blood.

The reality is harsh for Intelers, no matter how hard they worked for that organization, when they become old by Intel standards, they are dumped. With a shrinking market share, more and more Intel people will join FACE Intel.

Wednesday, September 19, 2007

Intel's produly displays answer to Native Quad-Core

Non-native is beautiful, says Intel. Sigh!

Intel engineers have the thickest skin, putting their primitive, crude, retarded, crappy, double cheeseburger design on their back. It's like saying -- "see mom, what an Intel moron".

Even Ratty says they need some originality.

Intel will do tri-core, following AMD's behind

Intel's tri-core processor will be three single cores strapped on a FSB.

One Intel engineer bragged that he can do it in no time.

The same Intel engineer claimed that he can do a 16-core chip if his boss wants it -- just strap 16 single cores on the FSB -- no brainer. As for benchmarks, just choose those which run in cache - 16 copies of super-PI will get very good results and make Intel fanbois dance.

In other news, Intel announced that it will have something called quick-path, a knock-off HyperTransport link in 2009 time frame. Dirk Meyer is laughing at the copycats.

Also, Intel will finally have an IMC, another idea copied from AMD. Intel says they had an IMC with their 8088, and they had an idea similar to AMD Fusion in 1990.

Jerry Sanders had those ideas back when he was at Fairchild Semi.

Tuesday, September 18, 2007

AMD should do a 1.5 core CPU

With K10's advanced technology of clocking core independently, what I proposed all along can be easily realized -- an asymmetric processor with two cores running at different speeds.

It will be very nice to have one core running at 3.6GHZ and the other at 2GHZ.

Intel has 1.5 core technology already. An Intel quad core is more like 1.5 core. 4 cores on a shared FSB is like 4 dogs breathing through a single straw -- a retarded situation.

Friday, September 14, 2007

Lucas: Intel = Evil Dark Side

See this report. Even I am surprise to see an executive from Lucas Films to call Intel evil.

Intel pratices exlcusionism, which is a form of national socialism.

Sunday, September 09, 2007

AMD crushes Intel when using industry standard compiler

GCC is the world's most widely used compiler. GCC is available for virtually all platforms.

With GCC, Intel is simply fragged to pieces.

Some more benchmarks: 2.5GHZ AMD is 40% faster than 3GHZ Intel. It is K8 over Netbust all over again, or even more so. See this performance benchmark of two equally clocked CPUs.

AMD sold 15,000 Barcelona CPUs to one customer already. FAB36 is cranking.

Patty feels dumb and Ratty is mad at his engineers.

Benchmarks show Opteron QC 1.9GHZ frags Clovertown 2.66GHZ

K10 1.9GHZ . Clovertown 2.66GHZ.

Intel has found it pretty hard to hide K10's massive 50% performance advantage. We expect Intel to immediately cut its prices by 50% or more, the 3GHZ Clovertown may be sold at $158, just below the 1.7GHZ Opteron QC.

Even Anand had to concede that K10 is fast. It is obvious that K10 has more than 20% integer IPC advantage over Intel.

Saturday, September 08, 2007

Randy Allen: Clovertown looks VERY anemic (emphasis added)

Randy Allen starts the talking. Hector is smiling.

Patty is polishing his resumes -- for different potential jobs.

Dr. Rattner is shouting at his stupid engineers and demanding for some originality. The Israeli team lost a lot of productivity because of engagement with Hezbollah last year -- they could have made Pentium 3 even better, you know. All Intel does now is adding even more cache.

Rumor says the Core 2 Duo on this warship failed when Hezbollah launched a C-801 at it.

Intel's only trick was cache for benchmarketing

A message posted by a reader at ZD-NET. AMD is expected to introduce true quad core on Monday, Intel is expected to follow suit in 2009.

/////////////////////////////QUOTE///////////////////////////////////
You hit the nail on the head
I assume you mistyped "7350" and meant "7320"... If that's the case then, yes, you hit the nail on the head: the only reason the 7345 beats the 7320 is because of its bigger L2 cache.

- Xeon 7320 (2.13 GHz, 4 MB L2): SPECint_rate2006=139
- Xeon 7345 (1.86 GHz, 8 MB L2): SPECint_rate2006=166

The 7345 clock frequency is 13% lower but its score is 19% higher. To better evaluate the impact of the cache size, I went ahead and compared 2 similarly clocked Tigertons:

- Xeon 7330 (2.40 GHz, 6 MB L2): SPECint_rate2006=178 [1]
- Xeon 7340 (2.40 GHz, 8 MB L2): SPECint_rate2006=192 [2]

A jump from 6 MB to 8 MB increases the score by 8%. Here is a breakdown of the scores obtained by each of the 12 benchmarks composing SPECint_rate2006:

1.187 bzip2 - Compression
1.178 hmmer - Search Gene Sequence
1.100 omnetpp - Discrete Event Simulation
1.097 mcf - Combinatorial Optimization
1.094 xalancbmk - XML Processing
1.089 libquantum - Physics / Quantum Computing
1.083 gcc - C Compiler
1.043 astar - Path-finding Algorithms
1.041 h264ref - Video Compression
1.019 perlbench - Programming Language
1.008 gobmk - Artificial Intelligence: Go
1.004 sjeng - Artificial Intelligence: Chess

(For example: "1.187 bzip2" means that the 7340 scores 18.7% better than the 7330 in the bzip2 benchmark: 146/123=1.187 according to [1] and [2].)

The increased cache size benefits mostly 2 of the SPECint_rate2006 benchmarks by almost 20%: bzip2 and hmmer.

Of particular importance is also the negligible impact of the cache size on h264ref. It suggests Tigerton's big L2 cache will almost be of no help when competing against K10 on video encoding tasks. I am sure power users will notice ;-)

Despite its modest L2/L3 cache sizes and given the generally impressive results K10 obtains at equal clock speeds against Core2, I can't wait to see what the 3.0 GHz Phenom monster is capable of.

[1] http://www.spec.org/cpu2006/results/res2007q3/cpu2006-20070821-01882.html
[2] http://www.spec.org/cpu2006/results/res2007q3/cpu2006-20070821-01881.html
///////////////////////////////END///////////////////////////////

AMD follows Intel on power specification

AMD's 95 watts is actually 75 in Intel's terms.

Intel follows the footsteps of AMD on everything else.

2GHZ K10 will frag 3GHZ Clovertown in most benchmarks.

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.

Thursday, September 06, 2007

AMD fabs are cranking day and night



FAB36 is fully equipped for K10 production. Those $200 Barcelona CPUs make people happy - except Intelers.

Intel's yield is so low, they can't make native quad core due to high defect density. Intel admits this. It has to wait for 45nm to get reasonable quad core output. Currently, Intel quads are just double cheese burger design -- zero originality. Intel CTO, Dr. Rattner, feels ashamed by Intel engineering. A couple of months ago, a top Intel CPU architect commented that Barcelona is too advanced to be feasible. His logic was that since K10 is too advanced even for Intel, AMD cannot possibly deliver it.



AMD's K10 yield is above 90%, due to the wonders of APM.

Wednesday, September 05, 2007

Americans and Chinese are smarter

AMD has over 50% of the retail market in both US and China.

Third world nations buy Intel.



Tuesday, September 04, 2007

Decent eight way computing for $1620

Two AMD K10 quad core CPUs (1.7GHZ 68 watt): $412
Tyan 1U rack server with 4x SATA: $622.
4GB of RAM: $310
4x 250 GB HDD: $276

Sunday, September 02, 2007

K10 demo video

2.0GHZ K10 30% faster than 2.33GHZ Clovertown (78/60).

The 2GHZ K10 is faster enough to frag a 3GHZ Intel quad.

But performance is not the only story, K10 has so many advanced features, it will take Intel years to even understand them.

Someone says that Intel engineers and scientists lack creativity.

Indeed, Core 2 looks so silly when compared to the K10. Intel dudes get a C- on CPU design from their own CTO. Who was that Israeli dude bragging about 40% lead over K8? I told you they were just patching up Bob Colwell's P6.