Poster | Thread |
wawa
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 10:20:52
| | [ #61 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| @Seiya
sundown wrote:
Quote:
Debian Squeeze has no h/w acceleration, only s/w. I ran glxinfo in a terminal to test. Not much point in posting benchmarks, they would just be slow because of it. Blender times in linux vs os4.1 with 1 core are the same, neither has h/w acceleration yet. |
Quote:
That wasn't my point anyway, all benchmarks are optimized for x86, some versions of linux might support h/w acceleration, I'm not a linux expert/fan to know. There are no fair tests for the x1000, some ppl here understand that, others use the results to degrade a-eon's efforts. |
Quote:
The x1000 runs os4 faster then any x86 system build, thats all I care about. |
Tommysammy:
Quote:
Yes right ,at this moment we don`t have any hardware2D and 3D driver. We have only 2D and 3D softwaredriver,so we don`t need to make any benchmarktests |
at least we can literally see with what kind of explanation templates the customers and testers are fed.Last edited by wawa on 03-May-2012 at 10:23 AM. Last edited by wawa on 03-May-2012 at 10:21 AM.
|
|
Status: Offline |
|
|
KimmoK
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 11:13:56
| | [ #62 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2003 Posts: 5211
From: Ylikiiminki, Finland | | |
|
| @itix
"To me interesting is "what if morphos was running on this hardware?" and these benchmarks answer my question very well."
It would run as fast as on the fastest G4 PowerMac? And it does not indicate that it is worthwhile for MOS to go (single corely) to G5 or PA6T? Especially performance wise not interesting for you to spend those thousands of euros for another PPC toy?
For me... x1000 is neither for me. (even though it's a very cool HW) My AOS4 HW investments will go to some other high performing PPC HW. (in the meanwhile I should get extra bang by registering MOS for my MosMini and re-install AROS to my old 2500+ AMD box.)
btw. would you consider AOS4 powered LimeBook as a interesting toy, even though you have plenty of G4PowerBooks(I imagine)? Last edited by KimmoK on 03-May-2012 at 11:15 AM.
_________________ - KimmoK // For freedom, for honor, for AMIGA // // Thing that I should find more time for: CC64 - 64bit Community Computer? |
|
Status: Offline |
|
|
pavlor
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 14:20:10
| | [ #63 ] |
|
|
 |
Elite Member  |
Joined: 10-Jul-2005 Posts: 9660
From: Unknown | | |
|
| @itix
Quote:
Wow, Honeywell, now that company has got bad reputation in Finland... |
I hope not for PA6T benchmarks. 
I linked that paper because it also confirms very good FFT performance of PA6T in comparison to G4 - same as we saw in FFT benchmark in HardInfo.
Quote:
FFTW does not efficiently exploit both processor cores, limiting |
Improtant part is:
Quote:
Similar trends observed on 8641D (and 970FX SMP with 2 processors) |
|
|
Status: Offline |
|
|
tommysammy
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 14:28:39
| | [ #64 ] |
|
|
 |
Cult Member  |
Joined: 20-Jan-2010 Posts: 664
From: Isselburg,Germany | | |
|
| @sundown
Quote:
sundown wrote: @tommysammy
I was hoping to try Debian Mint soon, it might have a better gfx driver? |
I don`t know it but i hope it_________________ Amiga600/Vampire2/PrismaMegaMix |
|
Status: Offline |
|
|
pavlor
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 14:33:22
| | [ #65 ] |
|
|
 |
Elite Member  |
Joined: 10-Jul-2005 Posts: 9660
From: Unknown | | |
|
| @KimmoK
Quote:
Only surprice to me is the higher FPU performance of the G5. Why is that? |
970MP 2500 MHz has SpecFp2000 result 2119. That is in the league of Pentium 4 Extreme Edition 3.73 GHz.
Quote:
So the 25W PA6T (and G4) seems very nice against 100W G5. |
HardInfo probably isn´t good benchmark for overall CPU performance.
Now comparison to 970MP in HardInfo benchmarks:
PA6T 1.8 GHz is as fast as 970 MP: CPU Blowfish: 1.67 GHz CPU CryptoHash: 2.16 GHz CPU Fibonacci: 2.6 GHz CPU N-Queens: 1.86 GHz FPU FFT: 1.50 GHz FPU Raytracing: 1.10 GHzLast edited by pavlor on 03-May-2012 at 02:39 PM.
|
|
Status: Offline |
|
|
andres
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 15:13:28
| | [ #66 ] |
|
|
 |
Regular Member  |
Joined: 3-Nov-2008 Posts: 329
From: Firenze (Italy) | | |
|
| Quote:
PA6T 1.8 GHz is as fast as 970 MP: CPU Blowfish: 1.67 GHz CPU CryptoHash: 2.16 GHz CPU Fibonacci: 2.6 GHz CPU N-Queens: 1.86 GHz FPU FFT: 1.50 GHz FPU Raytracing: 1.10 GHz |
If I understand well it seems PA6T is quite similar to a G5 in these performances...Last edited by andres on 03-May-2012 at 03:13 PM.
_________________ A1200/020+68882 - 6 MB RAM - AmigaOS 3.0
Home Recording Audio |
|
Status: Offline |
|
|
KimmoK
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 15:46:55
| | [ #67 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2003 Posts: 5211
From: Ylikiiminki, Finland | | |
|
| @pavlor
>970MP 2500 MHz has SpecFp2000 result 2119.
And as it seems PA6T does not seem to deliver according it's predicted SpecFP rating... _________________ - KimmoK // For freedom, for honor, for AMIGA // // Thing that I should find more time for: CC64 - 64bit Community Computer? |
|
Status: Offline |
|
|
amigadave
 |  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 15:59:07
| | [ #68 ] |
|
|
 |
Super Member  |
Joined: 18-Jul-2005 Posts: 1732
From: Lake Shastina, Northern Calif. | | |
|
| @andres
I don't know if any of these benchmarks are accurately showing the true performance of both cores in the PA6T, but if the test results eventually show that the 25watt PA6T is on par with a 100watt G5 of the same clock speed, that is a good accomplishment, to reduce power consumption by 75% while keeping the same processing power. _________________ Amiga! The computer that inspired so many, to accomplish so much, but has ended up in the hands of . . . . . . . . . . |
|
Status: Offline |
|
|
pavlor
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 16:26:49
| | [ #69 ] |
|
|
 |
Elite Member  |
Joined: 10-Jul-2005 Posts: 9660
From: Unknown | | |
|
| @KimmoK
Quote:
And as it seems PA6T does not seem to deliver according it's predicted SpecFP rating... |
Judging by HardInfo benchmarks it nearly delivers promised (PASemi expected 1500 SpecFp2000 at 2 GHz - 1350 at 1.8 GHz). However, other benchmarks I trust more (Blender) show much lesser results. |
|
Status: Offline |
|
|
A3000T
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 17:37:39
| | [ #70 ] |
|
|
 |
Cult Member  |
Joined: 7-Nov-2003 Posts: 633
From: the Netherlands | | |
|
| @amigadave
Quote:
I don't know if any of these benchmarks are accurately showing the true performance of both cores in the PA6T, but if the test results eventually show that the 25watt PA6T is on par with a 100watt G5 of the same clock speed, that is a good accomplishment, to reduce power consumption by 75% while keeping the same processing power. |
The PA6T uses only 7Watt maximum at 2GHZ. Not 25. The PA6T was designed to have a good performance/Watt ratio.
Kind regards,
Dennis
|
|
Status: Offline |
|
|
sundown
 |  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 17:43:53
| | [ #71 ] |
|
|
 |
Elite Member  |
Joined: 30-Aug-2003 Posts: 5120
From: Right here... | | |
|
| @Jupp3
Quote:
Does blender ever use h/w acceleration for rendering anyway? |
Good question, I have a feeling the answer is no.
I got 6:54 min with Wazp3D/os4.1 Linux 1 core 7 min Linux 2 cores 3:26 min
Not as fast as I would expect compared to my micro._________________ Hate tends to make you look stupid... |
|
Status: Offline |
|
|
pavlor
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 18:01:55
| | [ #72 ] |
|
|
 |
Elite Member  |
Joined: 10-Jul-2005 Posts: 9660
From: Unknown | | |
|
| @A3000T
Quote:
The PA6T uses only 7Watt maximum at 2GHZ. Not 25. The PA6T was designed to have a good performance/Watt ratio. |
One PA6T core 7 W. Two cores 14 W. Entire SoC 25 W. |
|
Status: Offline |
|
|
pavlor
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 18:03:33
| | [ #73 ] |
|
|
 |
Elite Member  |
Joined: 10-Jul-2005 Posts: 9660
From: Unknown | | |
|
| @sundown
Quote:
It would be really nice to have support for two cores even in AmigaOS.
Quote:
Not as fast as I would expect compared to my micro. |
Under 10 mins, right? |
|
Status: Offline |
|
|
sundown
 |  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 18:59:16
| | [ #74 ] |
|
|
 |
Elite Member  |
Joined: 30-Aug-2003 Posts: 5120
From: Right here... | | |
|
| @pavlor
10:42 min on my 800MHz micro, so I would expect under 5 min on the x1000 with 1 core.
_________________ Hate tends to make you look stupid... |
|
Status: Offline |
|
|
michalsc
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 19:18:28
| | [ #75 ] |
|
|
 |
AROS Core Developer  |
Joined: 14-Jun-2005 Posts: 421
From: Germany | | |
|
| @sundown
what scares me is, the PPC architecture used to be faster in that kind of math long time ago and it's not, anymore. An i5 cpu at 3GHz needs 25 seconds to render that scene in one thread. 7.22 seconds in four threads.
scary... |
|
Status: Offline |
|
|
tibby.dude
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 19:33:14
| | [ #76 ] |
|
|
 |
New Member |
Joined: 3-Jan-2012 Posts: 7
From: Cape Town | | |
|
| Okay here is some test.blend on my G5 running Blender 249.2
3:43.06 (1 threads) 1:59:46 (2 threads)
|
|
Status: Offline |
|
|
pavlor
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 19:38:30
| | [ #77 ] |
|
|
 |
Elite Member  |
Joined: 10-Jul-2005 Posts: 9660
From: Unknown | | |
|
| @tibby.dude
Quote:
Quote:
Nice!
However, not directly comparable to 2.48 results of OS4 Blender. |
|
Status: Offline |
|
|
broadblues
 |  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 20:16:37
| | [ #78 ] |
|
|
 |
Amiga Developer Team  |
Joined: 20-Jul-2004 Posts: 4449
From: Portsmouth England | | |
|
| @sundown
Quote:
Quote:
Does blender ever use h/w acceleration for rendering anyway?
|
Good question, I have a feeling the answer is no.
|
Blender 2.4x and blender 2.5x (for low x at anyrate)do not use the GPU and definetly not GL for rendering with the blender render engine.
faster Graphics would make a *small* difference if rendering with GUI open. Probably not enough to worry about though. GL inc shaders *is* used for rendering with the game engine.
For very much later blenders 2.5x (high x) and 2.6x there is anew engine called cycles whch uses the GPU (not GL) via nVidia;s cuda (not 100% sure I got the acronym right) interface to the GPU.
Quote:
I got 6:54 min with Wazp3D/os4.1 Linux 1 core 7 min Linux 2 cores 3:26 min
|
Those figures don't look to bad to me. It's quite aheavy benchmark IRC, That's comarable to a 2.6 GHz pentiom 4 whatever that means in real life.
Bear in mind also that CPU grunt isn't the only limiting factore with blender, very complex scenes need memory. Sometimes quite alot of it, so the X1000 is better placed than a micro in that regard.
_________________ BroadBlues On Blues BroadBlues On Amiga Walker Broad |
|
Status: Offline |
|
|
wawa
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 20:17:21
| | [ #79 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| @sundown
Quote:
by sundown on 3-May-2012 18:43:53 @Jupp3 Quote: Does blender ever use h/w acceleration for rendering anyway? Good question, I have a feeling the answer is no. |
good thing you start slowly to realize. to my knowledge 3d software does not use hardware acceleration for rendering because the results might vary too much depending on used hardware. it might be used for preview instead.
@broadblues: you ve beat me, but thanks for confirmation..;)Last edited by wawa on 03-May-2012 at 08:18 PM.
|
|
Status: Offline |
|
|
wawa
|  |
Re: Debian Squeeze Hardinfo Benchmarks Posted on 3-May-2012 20:21:13
| | [ #80 ] |
|
|
 |
Elite Member  |
Joined: 21-Jan-2008 Posts: 6259
From: Unknown | | |
|
| @pavlor
Quote:
However, not directly comparable to 2.48 results of OS4 Blender. |
so you expect a major blender engine overhaul somewhere between 2.48 and 2.49 that might have significant influence on the benchmark? lets say 200% speedup? |
|
Status: Offline |
|
|