Tuesday, February 24, 2009

Safari 4 - fastest ? ! IE8 RC nowhere close its slowest

So now that Safari 4 beta has come , i know its supposed to be the fastest, so i am seeing if google chrome 2 which is in pre-beta breaks safari 4 beta. I used Safari4 beta today and i can see it is really fast.

Sad that though i am using IE8 RC, its so slow that its sad to see the gap, its nowhere close to Safari4. Chrome, Firefox (though firefox new version which is in beta will be also very fast), and also Opera.

I have done some benchmarking, see the comparision its surprising how far away is IE8
This is on Windows 7


TEST COMPARISON IE8 RC Chrome 2.0.164 (nightly) DETAILS

=============================================================================

** TOTAL **: 5.83x as fast 12379.2ms +/- 0.8% 2122.6ms +/- 4.0% significant

=============================================================================

3d: 4.49x as fast 1370.4ms +/- 1.8% 305.0ms +/- 14.1% significant
cube: 6.37x as fast 473.6ms +/- 4.9% 74.4ms +/- 16.7% significant
morph: 3.29x as fast 411.4ms +/- 2.8% 125.0ms +/- 28.8% significant
raytrace: 4.60x as fast 485.4ms +/- 3.8% 105.6ms +/- 15.9% significant

access: 9.59x as fast 2042.2ms +/- 1.3% 213.0ms +/- 8.6% significant
binary-trees: 18.3x as fast 329.4ms +/- 4.1% 18.0ms +/- 19.5% significant
fannkuch: 13.1x as fast 867.6ms +/- 0.8% 66.4ms +/- 14.8% significant
nbody: 6.19x as fast 494.8ms +/- 2.3% 80.0ms +/- 10.1% significant
nsieve: 7.21x as fast 350.4ms +/- 1.2% 48.6ms +/- 26.5% significant

bitops: 9.17x as fast 1731.2ms +/- 3.8% 188.8ms +/- 14.3% significant
3bit-bits-in-byte: 17.7x as fast 283.8ms +/- 17.4% 16.0ms +/- 18.2% significant
bits-in-byte: 10.1x as fast 309.8ms +/- 4.4% 30.6ms +/- 2.2% significant
bitwise-and: 10.1x as fast 637.8ms +/- 9.4% 63.0ms +/- 10.7% significant
nsieve-bits: 6.31x as fast 499.8ms +/- 3.8% 79.2ms +/- 28.0% significant

controlflow: 27.3x as fast 256.8ms +/- 1.1% 9.4ms +/- 15.1% significant
recursive: 27.3x as fast 256.8ms +/- 1.1% 9.4ms +/- 15.1% significant

crypto: 6.11x as fast 813.6ms +/- 4.5% 133.2ms +/- 16.6% significant
aes: 9.26x as fast 372.2ms +/- 9.6% 40.2ms +/- 40.3% significant
md5: 4.16x as fast 222.8ms +/- 1.1% 53.6ms +/- 14.5% significant
sha1: 5.55x as fast 218.6ms +/- 1.3% 39.4ms +/- 41.9% significant

date: 2.82x as fast 823.6ms +/- 3.2% 291.8ms +/- 7.3% significant
format-tofte: 2.91x as fast 439.6ms +/- 5.8% 151.2ms +/- 8.2% significant
format-xparb: 2.73x as fast 384.0ms +/- 2.2% 140.6ms +/- 11.1% significant

math: 4.67x as fast 1169.2ms +/- 3.7% 250.2ms +/- 6.2% significant
cordic: 4.24x as fast 504.0ms +/- 3.6% 119.0ms +/- 14.3% significant
partial-sums: 4.06x as fast 342.8ms +/- 14.3% 84.4ms +/- 11.3% significant
spectral-norm: 6.89x as fast 322.4ms +/- 1.4% 46.8ms +/- 7.1% significant

regexp: 12.3x as fast 668.0ms +/- 6.7% 54.4ms +/- 46.0% significant
dna: 12.3x as fast 668.0ms +/- 6.7% 54.4ms +/- 46.0% significant

string: 5.18x as fast 3504.2ms +/- 1.5% 676.8ms +/- 2.8% significant
base64: 17.9x as fast 1762.0ms +/- 1.8% 98.6ms +/- 22.4% significant
fasta: 6.04x as fast 548.4ms +/- 5.1% 90.8ms +/- 21.4% significant
tagcloud: 3.80x as fast 448.0ms +/- 4.4% 118.0ms +/- 5.4% significant
unpack-code: 1.60x as fast 371.6ms +/- 4.0% 231.6ms +/- 7.9% significant
validate-input: 2.72x as fast 374.2ms +/- 7.6% 137.8ms +/- 11.2% significant


--- Now between Chrome 2.0.164 and Safari 4 beta

TEST                   COMPARISON            Chrome             Safari 4             DETAILS

=============================================================================

** TOTAL **: *1.45x as slow* 2122.6ms +/- 4.0% 3076.8ms +/- 12.5% significant

=============================================================================

3d: *2.28x as slow* 305.0ms +/- 14.1% 695.8ms +/- 19.8% significant
cube: *4.53x as slow* 74.4ms +/- 16.7% 337.4ms +/- 18.8% significant
morph: *1.53x as slow* 125.0ms +/- 28.8% 191.6ms +/- 19.7% significant
raytrace: *1.58x as slow* 105.6ms +/- 15.9% 166.8ms +/- 25.4% significant

access: *1.75x as slow* 213.0ms +/- 8.6% 373.6ms +/- 19.8% significant
binary-trees: *2.13x as slow* 18.0ms +/- 19.5% 38.4ms +/- 1.8% significant
fannkuch: 1.52x as fast 66.4ms +/- 14.8% 43.6ms +/- 31.4% significant
nbody: *3.30x as slow* 80.0ms +/- 10.1% 263.8ms +/- 24.6% significant
nsieve: 1.75x as fast 48.6ms +/- 26.5% 27.8ms +/- 3.7% significant

bitops: 1.28x as fast 188.8ms +/- 14.3% 147.8ms +/- 17.8% significant
3bit-bits-in-byte: 1.67x as fast 16.0ms +/- 18.2% 9.6ms +/- 26.9% significant
bits-in-byte: 1.59x as fast 30.6ms +/- 2.2% 19.2ms +/- 39.8% significant
bitwise-and: 2.22x as fast 63.0ms +/- 10.7% 28.4ms +/- 2.4% significant
nsieve-bits: ?? 79.2ms +/- 28.0% 90.6ms +/- 28.0% not conclusive: might be *1.14x as slow*

controlflow: ?? 9.4ms +/- 15.1% 11.2ms +/- 42.6% not conclusive: might be *1.19x as slow*
recursive: ?? 9.4ms +/- 15.1% 11.2ms +/- 42.6% not conclusive: might be *1.19x as slow*

crypto: *1.71x as slow* 133.2ms +/- 16.6% 228.4ms +/- 10.6% significant
aes: ?? 40.2ms +/- 40.3% 45.8ms +/- 17.3% not conclusive: might be *1.14x as slow*
md5: *1.76x as slow* 53.6ms +/- 14.5% 94.4ms +/- 22.5% significant
sha1: *2.24x as slow* 39.4ms +/- 41.9% 88.2ms +/- 20.3% significant

date: 1.14x as fast 291.8ms +/- 7.3% 256.8ms +/- 12.6% significant
format-tofte: 1.72x as fast 151.2ms +/- 8.2% 88.0ms +/- 9.0% significant
format-xparb: *1.20x as slow* 140.6ms +/- 11.1% 168.8ms +/- 19.3% significant

math: *2.20x as slow* 250.2ms +/- 6.2% 549.4ms +/- 13.9% significant
cordic: *1.91x as slow* 119.0ms +/- 14.3% 226.8ms +/- 20.0% significant
partial-sums: *2.59x as slow* 84.4ms +/- 11.3% 218.8ms +/- 7.3% significant
spectral-norm: *2.22x as slow* 46.8ms +/- 7.1% 103.8ms +/- 18.6% significant

regexp: ?? 54.4ms +/- 46.0% 77.2ms +/- 25.0% not conclusive: might be *1.42x as slow*
dna: ?? 54.4ms +/- 46.0% 77.2ms +/- 25.0% not conclusive: might be *1.42x as slow*

string: *1.09x as slow* 676.8ms +/- 2.8% 736.6ms +/- 8.5% significant
base64: - 98.6ms +/- 22.4% 86.8ms +/- 21.2%
fasta: *1.36x as slow* 90.8ms +/- 21.4% 123.4ms +/- 4.2% significant
tagcloud: *1.90x as slow* 118.0ms +/- 5.4% 224.0ms +/- 3.2% significant
unpack-code: 1.52x as fast 231.6ms +/- 7.9% 152.8ms +/- 11.2% significant
validate-input: ?? 137.8ms +/- 11.2% 149.6ms +/- 26.3% not conclusive: might be *1.09x as slow*

Monday, February 23, 2009

Microsoft training ? Am I am PC ?

how sweet you think isnt it ... By chance are the going to teach linux internals and java ? And if they teach .NET and Visual Studio, i will consider it wrong, since thats forcing technological choice on people. its like the kids saying i am a PC in the Ads !

People should be taught computer science independent of vendors like Microsoft or Apple or Linux. I would be critical if it is a way of promoting microsoft products in the guise of training.

I am 1 year old and i am a PC ! how much impact does that make ?

Sunday, February 22, 2009

Life with windows 7

Well its been long since i have been using windows 7, though i still wonder why there are times when if something is not responding, it leads to a total chaos, and the whole system fails to respond, and the only way to fix it is hard booting.

I guess the killing process is screwed up somewhere. It seems like they are trying to restart properly, sort of seamlessly but that isnt working that well. Also this dwm.exe and explorer.exe, sometimes they grow and shrink on their own, i dont know whats with those two guys, trying to control the look and feel they are having a tough time.

Now getting back to features, well to be true, its a natural progression from windows vista, which was really badly recieved. Also with the fact that OSX doesnt allow legal installations on PC's , and no real big corporations planning and marketing linux, we are having nothing else. I mean there is nothing that can kill windows, even if they release crap, there are people who will feed off off it.

The true competition would be when
1. OSX is also allowed on PC's and Netbooks.
2. Ubuntu is aggresively marketed.
3. Japanese companies concentrate on building not just hardware but also software (OS i mean).
4. IBM aggresively markets and creates AIX for desktops and servers.
5. Sun solaris.
6. Google android for PC.

I guess an OS that universally works on all platforms, would be a true OS. Google has a better strategy since they built a modified version of linux for mobiles, that made them think of conserving space, and if they are able to use that OS with little modification onto PC's we have a great OS with a lower memory footprint.

Slumdog wins !

Well interestingly Slumdog won so many oscars ! my god !