Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: ikir on March 15, 2003, 01:31:35 PM
-
On IkS (http://www.iksnet.it) you can find some new AmigaONE test done by Massimiliano Tretene (Soft3)
-------------------------
Dalla mailing-list relativa all'os4:
"Sto testando in questi giorni un nuovo dnet client con ottimizzazioni per l'altivec.
Da test preliminari, il core rc5-72 ottimizzato altivec, raggiunge gli 8 Megakeys / sec.
Questo dato è da confrontare con i 4.65 Megakeys / sec raggiunti da un Athlon ad 1.8 Ghz.
Sono d'accordo che questo genere di test sia molto generico, però può dare una indicazione approssimativa su che tipo di accelerazione si può attendere da una applicazione che tragga beneficio dall'utilizzo dell'altivec.
Massimiliano Tretene"
--------------------------
Its about altivec. It seem that AmigaONE+Altivec give a double speed than a Athlon1.8
-
I'm thinking about a game-killer applications wich runs with Altivec Support......
-
"From the mailing-list relative to the os4: "I am testing in these days a new one dnet client with optimizations for the altivec. From preliminary tests, optimized Core rc5-72 altivec, catches up the 8 Megakeys/sec. This data? to confront with the 4,65 Megakeys/sec caught up from a Athlon to 1.8 Ghz. They are of agreement that this kind of test is much generic one, for? pu? to give one approximate indication on that type of acceleration pu? to attend from an application that it draws benefit from I use of the altivec. Massimiliano Marked Tretene "from MDI Gives impressive said to you. We wait for to you on our furum in order to discuss about these tests."
I love computer translations! :-D
-
Eh eh! Great Elektro :-D
-
We want more! We want more! :-P
-
ksk wrote:
We want more! We want more! :-P
LOL 8-)
-
ksk wrote:
We want more! We want more! :-P
Here (http://n0cgi.distributed.net/speed/) you go!
(Maybe you were being sarcastic, it's hard to tell when XOOPS mistranslates a ":-P" smiley into a "happily grinning mongoloid" GIF... ;) )
A dnetc run is as good as entirely CPU-bound. It doesn't matter much if the CPU sits in a Teron or a Mac, or if it's running Linux or whatever.
For a system benchmark, it's almost entirely worthless, except for perhaps finding out if there are serious performance problems.
-
hasnt this been talked to death? I thought people understood by now.
that altivec is just a SIMD... and yes it's a highly optomiosed SIMD... but it's much better for improveing benchmarks then actually doing anything practical.
e.g running Apps.
the Athlon survives on it's real world capabilities, as does the P4... not on some benchmarks..
-
I agree. The G4 looks spectacular on paper, and
yes it's pretty fast, but there's something
wrong when dual G4's are getting beaten by
single x86 processors in the real - world
(applications) benchmarks. It wouldn't be
such a big deal if it weren't for the
cost/performance ratio.
The G4 is still a great processor though. My
sister's G4/700mhz feels pretty fast, especially
under OS 9. OS X seems pretty sluggish still
no matter how fast of a mac it's run on.
Hopefully that will change soon.
-
mips_proc wrote:
...
that altivec is just a SIMD... and yes it's a highly optomiosed SIMD... but it's much better for improveing benchmarks then actually doing anything practical.
e.g running Apps.
Word processing probably has no use for SIMD, but then again, it doesn't have much use for MHz either. On the other hand, DSP and image processing type applications can get a v. nice boost from it.
Problem is, you have to compile the apps to use it. Otherwise it's like having a floating point unit, but running apps compiled for an integer only prior version of the cpu.
A lot of stuff you may run on a PIV these days is compiled for PII, and you just don't get the kind of oomph you'd get if the code was compiled for the machine it ran on.
Hopefully, apps that could benefit from SIMD will ship with an Altivec optimized version as well as the G3 version, in the same way we get our 68000, 68020, and 68040 executables now.
-
What I really meant was that I/we want to see more benchmarks in general.
(But I'm also happy to see that the Altivec is useable in A1.)
-
i agree i want to see more A1 comparisons too as im waiting nfor my A1 to come it gives us somthing to look forward too as we dont know when our boards will come so please all you lucky ppl out there send up your speed tests for us to drewl over with altivec active
-
ikir wrote:
On IkS (http://www.iksnet.it) you can find some new AmigaONE test done by Massimiliano Tretene (Soft3)
-------------------------
Dalla mailing-list relativa all'os4:
"Sto testando in questi giorni un nuovo dnet client con ottimizzazioni per l'altivec.
Da test preliminari, il core rc5-72 ottimizzato altivec, raggiunge gli 8 Megakeys / sec.
Questo dato è da confrontare con i 4.65 Megakeys / sec raggiunti da un Athlon ad 1.8 Ghz.
Sono d'accordo che questo genere di test sia molto generico, però può dare una indicazione approssimativa su che tipo di accelerazione si può attendere da una applicazione che tragga beneficio dall'utilizzo dell'altivec.
Massimiliano Tretene"
--------------------------
Its about altivec. It seem that AmigaONE+Altivec give a double speed than a Athlon1.8
To illustrate the opposite view refer any OpenSSL benchmarks…
http://www.macosxhints.com/article.php?story=20020113045343563
Any good Linux distributions should have included OpenSSL utilities.
About OpenSSL utility, refer to
http://www.mkssoftware.com/docs/man1/openssl_x509.1.asp
-
WOW. Impressive. What is a megakey, BTW ? :-D
-
OS X 10.2.x is snappy provided you have a video card that supports QE (AGP GeForce2MX or better, AGP Radeon or better)
-
Well the OSX used that 'quartz' technology, which uses your grpahics card for desktop display.
-
Its about altivec. It seem that AmigaONE+Altivec give a double speed than a Athlon1.8
In that particular benchmark, yes..
But Distributed.net thinks that “the RC5 client is a poor benchmark to use in determining the speed or performance of a particular CPU.”
http://n0cgi.distributed.net/faq/cache/55.html
Couldn’t you a find benchmark that relies less on a certain type of instruction?
As for the reasons why G4 kicks butt in RC5 I would like to refer to a Slashdot IRC chat with the guys from distributed.net:
http://www.slashnet.org/forums/DCTI-20020928.html
“G4 CPUs have some architectural features very suited for RC5.”
“First, in the fastest cores, all processing is done in the vector unit of the chip (Altivec).”
“Intel and AMD CPUs do have integer vector units (SSE2 and MMX), but they're less suited to RC5 than Altivec for two main reasons:
More registers available (32 in the PowerPC versus 8 in MMX and SSE2), plus 128-bit wide registers (MMX is only 64-bit wide), and the existence of a hardware vector rotate instruction in Altivec, which isn't available in MMX and SSE2.”
“These reasons make it less worthy to use vector units on x86, where all processing is done in the standard scalar ALU.
Oh, and did I mention that the [G4] vector unit allows for the processing of 4 keys simultaneously?”
“The most recent AMD processors have better hardware rotate support than the most recent Intel ones, to answer the second question. RC5 uses rotate _a lot_”
“The AMD Athlon possesses 3 superscalar ALUs, all capable of doing a rotate instruction with single-cycle latency.
Given that rotates are the core of RC5 processing, this is very advantageous for the Athlon.
Also, the P4 latency for rotate instructions is 4 cycles, and while I'm not sure off the top of my head whether it has 1 or 2 superscalar barrel shifters, it is clearly in disadvantage here.”