Is LightWave using SSE on the Palomino?

G

Guest

Guest
I'm surprised to see the P4s beating the XPs so easily, so I'm wondering if it's using an equally clever method for CPU detection as Windows Media Encoder 7.

Anybody an idea what kind of code lightwave runs on non SSE2 or Altivec systems?
 

MeTaLrOcKeR

Distinguished
May 2, 2001
1,515
0
19,780
I think its safe to say it's non-AMD SEE optimized...just like Media Player =)

that's WAY to much of a beating that the P4 puts on the XP.....it's pure SSE2 optimized without the Strign to find the SSE Instructions on the Palomino.....i can bet a silver dollar on it...

-MeTaL RoCkEr

<A HREF="http://www.anandtech.com/mysystemrig.html?id=13597" target="_new"> <b> My Rig </b> </A>
 

mbetea

Distinguished
Aug 16, 2001
1,662
0
19,780
lw 7.0b is sse2 optimized as well. not sure how much more performance it would give exactly, but that could be the difference.

well if luck is a lady, it explains why i have no luck :frown:
 

MeTaLrOcKeR

Distinguished
May 2, 2001
1,515
0
19,780
It IS optimized for SSE and SSE2 YES...BUT i bet it doesnt detect the SSE on the Athlon XP becasue of the CPU_ID String nto Saying "Genuine Intel" .......

-MeTaL RoCkEr

<A HREF="http://www.anandtech.com/mysystemrig.html?id=13597" target="_new"> <b> My Rig </b> </A>
 

bront

Distinguished
Oct 16, 2001
2,122
0
19,780
I found it quite interesting to see the Celeron 1200 beating the P4s in a few benchmarks ("Lame" MPEG encoding, SiSoft CPU and Multimedia tests, Linux kernal compiling). You would think that their "Value" processors wouldn't be able to outperform their "Performance" processors, especialy in heavy computational issues.

Chesnuts roasting on an open CPU
Bill Gates nipping at your wallet
 
G

Guest

Guest
Well it *should* detect it, but I really think it doesn't (just like the famous Media Encoder), since Palomino and Tbird are effectivly scoring the same.

I don't have problems with a P4 beating the XPs per se, but IMVHO this benchmark is flawed.

PS: no "AMD zealot" flames please. Next system I buy for home use is very likely going to be a Tualaron 1.2. If I underclock and undervolt it I might just get away with passive cooling ;).
 

Copenhagen

Distinguished
Oct 21, 2001
552
0
18,980
DSTA wrote:

"I'm surprised to see the P4s beating the XPs so easily, so I'm wondering if it's using an equally clever method for CPU detection as Windows Media Encoder 7."

It's certainly a possibility, take a look here, notice that Intel is sponsering the Lightwave [7] World Tour !!

<A HREF="http://www.lightwave3d.com/intel/tour.html" target="_new">http://www.lightwave3d.com/intel/tour.html</A>

I have e-mail'ed the guys at Newtek which is responsible for Lightware 7b, and asked them if Lightwave 7b is able to utilize the AMD Athlon XP's support for SSE. I'll let you know as soon as I get a response.


In the meantime I think I'll acquire the program and start using it for no reason other than to feel good about my powerful P4. :))



--------------------- P4 1700 MHz ---------------------

Overcloked 1700 MHz P4 Willamette Socket 478 with stock cooling benchmarked using SiSoft Sandra. "Crippled" by an Antec 300W P4 power-supply.

CPU Arithmetic.
Dhrystone ALU:..............4039
Whetstone FPU/SSE2:.....1097/2571 MFLOPS

Multi-Media:
Integer iSSE2:...............8304 it/s
Floating-Point iSSE2:.....10109 it/s

Memory Bandwidth:
RAM Int MMX:...............1950 MB/s
RAM Float FPU:.............1957 MB/s
 

Copenhagen

Distinguished
Oct 21, 2001
552
0
18,980
There are two other possibilities that I can think of:

1.Intel made Newtek implement Lightwave 7b using only SSE2 instructions.

2.Intel made Newtek implement Lightwave 7b in a way which requires a massive memory bandwidth. This gives the P4 an advantage at least when combined with the Intel 850 chipset.

Of cause Intel may not have had any influence on Newtek's implentation at all, but it's more to do with the nature of the tasks that Lightwave has to carry out, i.e both SSE2 and memory bandwidth are important.
 
G

Guest

Guest
Thanks very much for writing Newtek, Copenhagen. I thought of doing the same, so you saved me some work ;).

Good points about LW7b only using SSE2 and being bandwith limited, but what could speak against that theory is that little Tualaron's score is comparable to 1.4 AMDs - this coming without SSE2 *and* having to live with the least mem bandwith.

Let's see if Newtek can shed some light on the matter.
 

AmdMELTDOWN

Distinguished
Dec 31, 2007
2,000
0
19,780
hello DSTA

Newtek Lightwave3d 7b takes advantage of SSE2 and at this
time the Palomino uses no SSE2, this is why lw is faster, there's
no surprises, dude! there are no nefarious deeds here,

also, Newtek optimizes for all platforms that'll will take
advantage of the program and vice versa, it only makes sense.

ps, don't be surprised or feel bad when SSE2 is actually
brilliantly implemented and it woops the AMD cpu with it's
3dnow, 3dnow pro, sse, mmx, did I forget one?

next: 3dmax w/SSE2

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
 

Matisaro

Splendid
Mar 23, 2001
6,737
0
25,780
Sure, im sure that intel giving them tons of cash didnt affect their code at all, how do you explaikn that in nearly every other rendering prog the xp kicks the p4s ass?

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
 
G

Guest

Guest
Oh I know Lightwave takes advantage of SSE2 on the PIV.

What I want to know is what happens on P3, Celeron and Athlon systems.
 

AmdMELTDOWN

Distinguished
Dec 31, 2007
2,000
0
19,780
>What I want to know is what happens on P3, Celeron and Athlon systems.

well, here help yourself <A HREF="http://www.blanos.com/benchmark/" target="_new">http://www.blanos.com/benchmark/</A>

matisaro, Newtek works with all platforms be it x86, Apple or DEC what does money have to do with SSE2 code optimizations? Is it me or you are acting like a baby, life's a bitch heh?

Ligthwave 7b just runs fast with SSE2 and Intel is touting the fact, it just makes sense! you ppl are just too stupid to get it.

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
 

mbetea

Distinguished
Aug 16, 2001
1,662
0
19,780
well, i'm shocked and going to have to pretty much agree with meltdown on this. high end apps like lightwave, maya, SI/XSI, after effects, combustion, etc. do not need to be swayed by intel. they might have pull with MS and the hardware relm, but i highly doubt their worried about whether intel is 100% behind them. they have their business elsewhere. where the pressure comes from is developers and customers. the people that these companies are making their money from are post houses, which EVERY and ANY advantage they can have with speeding up the process means cut's in manhours which leads to cuts in price. most people look at the benchmarks and see that the xp is a few points behind or so from the p4. but to SE supervisor, that's a big advantage.

well if luck is a lady, it explains why i have no luck :frown:
 

Matisaro

Splendid
Mar 23, 2001
6,737
0
25,780
Which is fine, I can accept if the sse2 is the reason ligfhtwave does so well on the p4. My question is why do you see the xp winning in other apps which perform the same tasks, and why intel just happens to be sponsoring the company who's apps perform so well with the p4.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!