HybridSLI works with OpenCL

J the Ninja · Aug 30, 2009, 03:33 AM

Well, no, not really. But you don't want it to.

First of all, let me say yet again, that "GeForce Boost", or running the 9400M and 9600M GT as an SLI pair is NOT supported in hardware. This is because the 9600M GT would have to scale down so much to match the 9400M, the resulting performance would be worse than the 9600M GT by itself.

Ok, first, the pure awesome news: OpenCL really and truly does support both cards at the same time.

More specifically, it sees them both as two entirely distinct devices, alongside the CPU. This is actually better than an SLI pair, as the software is free to decide which chip it wants to put to work for it's particular task. What follows is my own tests, from some tools posted over at the Mac Pro board. You can get them here:

(note that oclinfo requires registration at InsanelyMac for download. If someone has another link to it, or wants to start a torrent, post it here. That's ok to do, oclinfo is released under the GNU GPL)

http://www.macupdate.com/info.php/id...encl-benchmark

http://www.insanelymac.com/forum/ind...&#entry1177815

To install, unzip them and copy the executable file to your /usr/bin folder. (Use Finder's "Go to Folder" to get there, it will not appear normally. You will be prompted for your admin password). To run, simply open Terminal, type the name of the program and hit Enter.

Code:

1 OpenCL platform found!

[Platform 0]
	Name: 		Apple
	Vendor: 	Apple
	Version: 	OpenCL 1.0 (Jul 15 2009 23:07:32)
	Profile: 	FULL_PROFILE


[OpenCL-only Context]
3 OpenCL devices found!

[Device 0]
	Name: 			GeForce 9600M GT
	Vendor: 		NVIDIA
	Type: 			GPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	CLH 1.0
	Compute Units: 		32
	Work Group Size: 	512
	Clock: 			1250 MHz
	Global Memory: 		256 MB
	Local Memory: 		16 KB
	Cache Size: 		0 KB
	Cache Line Size: 	0 Bytes
	Available: 		Yes
	Double-Precision: 	No
	Extensions: 
				cl_khr_byte_addressable_store
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[Device 1]
	Name: 			GeForce 9400M
	Vendor: 		NVIDIA
	Type: 			GPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	CLH 1.0
	Compute Units: 		16
	Work Group Size: 	512
	Clock: 			1100 MHz
	Global Memory: 		256 MB
	Local Memory: 		16 KB
	Cache Size: 		0 KB
	Cache Line Size: 	0 Bytes
	Available: 		Yes
	Double-Precision: 	No
	Extensions: 
				cl_khr_byte_addressable_store
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[Device 2]
	Name: 			Intel(R) Core(TM)2 Duo CPU     P8600  @ 2.40GHz
	Vendor: 		Intel
	Type: 			CPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	1.0
	Compute Units: 		2
	Work Group Size: 	1
	Clock: 			2400 MHz
	Global Memory (Total): 	4096 MB
	Global Memory (Host): 	3072 MB
	Global Memory (PCIe): 	1024 MB
	Local Memory: 		16 KB
	Cache Size: 		3072 KB
	Cache Line Size: 	64 Bytes
	Available: 		Yes
	Double-Precision: 	Yes
	Extensions: 
				cl_khr_fp64
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_khr_local_int32_base_atomics
				cl_khr_local_int32_extended_atomics
				cl_khr_byte_addressable_store
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[shared OpenCL+OpenGL Context]
2 OpenCL devices found!

[Device 0]
	Name: 			GeForce 9600M GT
	Vendor: 		NVIDIA
	Type: 			GPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	CLH 1.0
	Compute Units: 		32
	Work Group Size: 	512
	Clock: 			1250 MHz
	Global Memory: 		256 MB
	Local Memory: 		16 KB
	Cache Size: 		0 KB
	Cache Line Size: 	0 Bytes
	Available: 		Yes
	Double-Precision: 	No
	Extensions: 
				cl_khr_byte_addressable_store
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[Device 1]
	Name: 			Intel(R) Core(TM)2 Duo CPU     P8600  @ 2.40GHz
	Vendor: 		Intel
	Type: 			CPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	1.0
	Compute Units: 		2
	Work Group Size: 	1
	Clock: 			2400 MHz
	Global Memory (Total): 	4096 MB
	Global Memory (Host): 	3072 MB
	Global Memory (PCIe): 	1024 MB
	Local Memory: 		16 KB
	Cache Size: 		3072 KB
	Cache Line Size: 	64 Bytes
	Available: 		Yes
	Double-Precision: 	Yes
	Extensions: 
				cl_khr_fp64
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_khr_local_int32_base_atomics
				cl_khr_local_int32_extended_atomics
				cl_khr_byte_addressable_store
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 3
OpenCL Device # 0 = GeForce 9600M GT
Device 0 is an: GPU with max. 1250 MHz and 32 units/cores 
Now computing - please be patient....
time used:  2.805 seconds

OpenCL Device # 1 = GeForce 9400M
Device 1 is an: GPU with max. 1100 MHz and 16 units/cores 
Now computing - please be patient....
time used:  3.081 seconds

OpenCL Device # 2 = Intel(R) Core(TM)2 Duo CPU     P8600  @ 2.40GHz
Device 2 is an: CPU with max. 2400 MHz and 2 units/cores 
Now computing - please be patient....
time used: 15.459 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

NOTE: When posting your results, PLEASE enclose it with the [code] tags, to make it easier to read!

As you can see here, the CPU, 9600M, and 9400M are ALL listed as devices! The two GPUs are available simultaneously, and the benchmark uses them one after the other. You will also notice, however, that when the device check is performed looking for combined OpenGL/OpenCL devices, the 9400 does not appear. As per usual, it is not available as an OpenGL device when the 9600 is running, and one more time...NEVER WILL BE. (I get sick of seeing Hybrid SLI threads here, if you couldn't tell.)

OK! Without further ado, let's get to benchmarking! Please post results from older (non-unibody) MacBook Pro's as well, and even other Apple laptops!

Also, sidenote: See how the 9600M GT is listed as 1.25Ghz? Meaning it isn't underclocked like some claim, at least not in it's max power state. It likely only underclocks at idle, or when running on battery.

Eddyisgreat · Aug 30, 2009, 04:00 AM

Macbook Pro "Classic" Early 2008 15" High

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce 8600M GT
Device 0 is an: GPU with max. 940 MHz and 32 units/cores 
Now computing - please be patient....
time used:  3.005 seconds

OpenCL Device # 1 = Intel(R) Core(TM)2 Duo CPU     T9300  @ 2.50GHz
Device 1 is an: CPU with max. 2500 MHz and 2 units/cores 
Now computing - please be patient....
time used: 15.304 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

DAMNiatx · 05:37 AM

Late 2008 MBP 2.4GHz
Snow Leopard Build 10a421
Open CL info

Code:

Last login: Sun Aug 30 16:34:57 on ttys000
Rheza-Pahlevis-MacBook-Pro:~ Rheza$ oclinfo
1 OpenCL platform found!

[Platform 0]
	Name: 		Apple
	Vendor: 	Apple
	Version: 	OpenCL 1.0 (Jul 15 2009 23:07:32)
	Profile: 	FULL_PROFILE


[OpenCL-only Context]
3 OpenCL devices found!

[Device 0]
	Name: 			GeForce 9600M GT
	Vendor: 		NVIDIA
	Type: 			GPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	CLH 1.0
	Compute Units: 		32
	Work Group Size: 	512
	Clock: 			1250 MHz
	Global Memory: 		256 MB
	Local Memory: 		16 KB
	Cache Size: 		0 KB
	Cache Line Size: 	0 Bytes
	Available: 		Yes
	Double-Precision: 	No
	Extensions: 
				cl_khr_byte_addressable_store
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[Device 1]
	Name: 			GeForce 9400M
	Vendor: 		NVIDIA
	Type: 			GPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	CLH 1.0
	Compute Units: 		16
	Work Group Size: 	512
	Clock: 			1100 MHz
	Global Memory: 		256 MB
	Local Memory: 		16 KB
	Cache Size: 		0 KB
	Cache Line Size: 	0 Bytes
	Available: 		Yes
	Double-Precision: 	No
	Extensions: 
				cl_khr_byte_addressable_store
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[Device 2]
	Name: 			Intel(R) Core(TM)2 Duo CPU     P8600  @ 2.40GHz
	Vendor: 		Intel
	Type: 			CPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	1.0
	Compute Units: 		2
	Work Group Size: 	1
	Clock: 			2400 MHz
	Global Memory (Total): 	4096 MB
	Global Memory (Host): 	3072 MB
	Global Memory (PCIe): 	1024 MB
	Local Memory: 		16 KB
	Cache Size: 		3072 KB
	Cache Line Size: 	64 Bytes
	Available: 		Yes
	Double-Precision: 	Yes
	Extensions: 
				cl_khr_fp64
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_khr_local_int32_base_atomics
				cl_khr_local_int32_extended_atomics
				cl_khr_byte_addressable_store
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[shared OpenCL+OpenGL Context]
2 OpenCL devices found!

[Device 0]
	Name: 			GeForce 9600M GT
	Vendor: 		NVIDIA
	Type: 			GPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	CLH 1.0
	Compute Units: 		32
	Work Group Size: 	512
	Clock: 			1250 MHz
	Global Memory: 		256 MB
	Local Memory: 		16 KB
	Cache Size: 		0 KB
	Cache Line Size: 	0 Bytes
	Available: 		Yes
	Double-Precision: 	No
	Extensions: 
				cl_khr_byte_addressable_store
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

[Device 1]
	Name: 			Intel(R) Core(TM)2 Duo CPU     P8600  @ 2.40GHz
	Vendor: 		Intel
	Type: 			CPU 
	Device Version: 	OpenCL 1.0
	Driver Version: 	1.0
	Compute Units: 		2
	Work Group Size: 	1
	Clock: 			2400 MHz
	Global Memory (Total): 	4096 MB
	Global Memory (Host): 	3072 MB
	Global Memory (PCIe): 	1024 MB
	Local Memory: 		16 KB
	Cache Size: 		3072 KB
	Cache Line Size: 	64 Bytes
	Available: 		Yes
	Double-Precision: 	Yes
	Extensions: 
				cl_khr_fp64
				cl_khr_global_int32_base_atomics
				cl_khr_global_int32_extended_atomics
				cl_khr_local_int32_base_atomics
				cl_khr_local_int32_extended_atomics
				cl_khr_byte_addressable_store
				cl_APPLE_gl_sharing
				cl_APPLE_SetMemObjectDestructor
				cl_APPLE_ContextLoggingFunctions

Rheza-Pahlevis-MacBook-Pro:~ Rheza$

opencl benchmark

Code:

Last login: Sun Aug 30 17:25:53 on ttys000
Rheza-Pahlevis-MacBook-Pro:~ Rheza$ opencl
...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 3
OpenCL Device # 0 = GeForce 9600M GT
Device 0 is an: GPU with max. 1250 MHz and 32 units/cores 
Now computing - please be patient....
time used:  2.794 seconds

OpenCL Device # 1 = GeForce 9400M
Device 1 is an: GPU with max. 1100 MHz and 16 units/cores 
Now computing - please be patient....
time used:  9.037 seconds

OpenCL Device # 2 = Intel(R) Core(TM)2 Duo CPU     P8600  @ 2.40GHz
Device 2 is an: CPU with max. 2400 MHz and 2 units/cores 
Now computing - please be patient....
time used: 18.394 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :) 
Rheza-Pahlevis-MacBook-Pro:~ Rheza$

DAMNiatx · Aug 30, 2009, 09:28 AM

hahahaha, no one care about it.
i know why apple need update their product more often.

jav6454 · Aug 30, 2009, 11:13 AM

It incredible how the 9600M GT could process things in ~2 seconds where as the CPU took ~10 seconds. I know he difference o cores makes a difference. But dam! Never knew it'd make that much.

Bearxor · Aug 30, 2009, 04:22 PM

2.4ghz MBP with a 256MB GF8600

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce 8600M GT
Device 0 is an: GPU with max. 940 MHz and 32 units/cores 
Now computing - please be patient....
time used:  2.930 seconds

OpenCL Device # 1 = Intel(R) Core(TM)2 Duo CPU     T7700  @ 2.40GHz
Device 1 is an: CPU with max. 2400 MHz and 2 units/cores 
Now computing - please be patient....
time used: 15.164 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

chris.niziolek · Aug 30, 2009, 05:25 PM

I get this when I run the utility.

Code:

(ComputerNAME):~ chris$ OpenCL2_Bench_V025
-bash: /usr/bin/OpenCL2_Bench_V025: Bad CPU type in executable
(ComputerName):~ chris$

Is it because I have an ATI Video Card?

Profiler Info Follows:

Code:

ATI Radeon X1600:

  Chipset Model:	ATY,RadeonX1600
  Type:	GPU
  Bus:	PCIe
  PCIe Lane Width:	x16
  VRAM (Total):	256 MB
  Vendor:	ATI (0x1002)
  Device ID:	0x71c5
  Revision ID:	0x0000
  EFI Driver Version:	01.00.068
  Displays:
Color LCD:
  Resolution:	1440 x 900
  Pixel Depth:	32-Bit Color (ARGB8888)
  Main Display:	Yes
  Mirror:	Off
  Online:	Yes
  Built-In:	Yes
AL2223W:
  Resolution:	1680 x 1050 @ 61 Hz
  Pixel Depth:	32-Bit Color (ARGB8888)
  Mirror:	Off
  Online:	Yes
  Rotation:	Supported

Computer Info Below

jav6454 · Aug 30, 2009, 06:01 PM

2.4GHz MacBook late 2008 with 4GB RAM. Seems my CPU is a slowpoke.

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce 9400M
Device 0 is an: GPU with max. 1100 MHz and 16 units/cores 
Now computing - please be patient....
time used:  3.510 seconds

OpenCL Device # 1 = Intel(R) Core(TM)2 Duo CPU     P8600  @ 2.40GHz
Device 1 is an: CPU with max. 2400 MHz and 2 units/cores 
Now computing - please be patient....
time used: 15.853 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :) 
logout

[Process completed]

Runt888 · Aug 30, 2009, 06:11 PM

Quote:

Originally Posted by J the Ninja

Ok, first, the pure awesome news: OpenCL really and truly does support both cards at the same time.

It seems like it only detects the 9600 if that's the card that is powering the display. So if you have Energy Saver set to the "Better battery life" setting, then the 9600 is completely shut off and not available for any computation (at least currently, although I would be willing to bet that will change if OpenCL really takes off).

tacojohn · Aug 30, 2009, 06:26 PM

24" iMac 2.8GHz 4GB RAM 8800 GS 512VRAM

Code:

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce 8800 GS
Device 0 is an: GPU with max. 1250 MHz and 64 units/cores 
Now computing - please be patient....
time used:  0.924 seconds

OpenCL Device # 1 = Intel(R) Core(TM)2 Duo CPU     E8235  @ 2.80GHz
Device 1 is an: CPU with max. 2800 MHz and 2 units/cores 
Now computing - please be patient....
time used: 12.855 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

jav6454 · Aug 30, 2009, 06:28 PM

Quote:

Originally Posted by tacojohn

24" iMac 2.8GHz 4GB RAM 8800 GS 512VRAM

Code:

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce 8800 GS
Device 0 is an: GPU with max. 1250 MHz and 64 units/cores 
Now computing - please be patient....
time used:  0.924 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

Dear God! not even 1 second!

King Mook Mook · Aug 30, 2009, 06:47 PM

Wow, this is really cool and I cannot wait until some developers start to make advantage of this! I tested this with my MacBook Pro 2.93 GHz, 4GB of RAM and the 9600 GT on with the 9400 not in use.

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 3
OpenCL Device # 0 = GeForce 9600M GT
Device 0 is an: GPU with max. 1250 MHz and 32 units/cores 
Now computing - please be patient....
time used:  2.794 seconds

OpenCL Device # 1 = GeForce 9400M
Device 1 is an: GPU with max. 1100 MHz and 16 units/cores 
Now computing - please be patient....
time used:  2.996 seconds

OpenCL Device # 2 = Intel(R) Core(TM)2 Duo CPU     T9800  @ 2.93GHz
Device 2 is an: CPU with max. 2930 MHz and 2 units/cores 
Now computing - please be patient....
time used: 13.932 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :) 
logout

[Process completed]

tacojohn · Aug 31, 2009, 12:19 AM

Quote:

Originally Posted by jav6454

Dear God! not even 1 second!

This is w/ 64-bit extensions active too.

jav6454 · Aug 31, 2009, 02:01 AM

Quote:

Originally Posted by tacojohn

This is w/ 64-bit extensions active too.

Mine was with 32-bit kernel. I want to do one with 64-bit to see if there really is a difference.

jpur · Aug 31, 2009, 02:35 AM

This were some of my runs from my June 2009 Macbook Pro. A couple of interesting things:

1. When I ran the test in 9400 mode, it did not recognize the 9600M

2. Running with 32 bit or 64 bit kernel made no difference

3. There was some variation in the 9400 benchmarks.

Still is really cool to see the raw power that future SL programs will be able to tap into. Thanks for sharing J!

glycon · Aug 31, 2009, 02:45 AM

From Apple's page on Concurrency and Application Design (http://developer.apple.com/mac/libra...8091-CH100-SW1).

The following Apple paragraphs seem a little contradictory, but this statement sums it up, "...any tasks that interact with the system are generally not recommended for use with OpenCL," in the second paragraph.

(relevant portion from Apple page follows)

OpenCL and Concurrency
The Open Computing Language (OpenCL) is a standards-based technology for performing general-purpose computations on a computer’s graphics processor. OpenCL is a good technology to use if you have a well-defined set of computations that you want to apply to large data sets. For example, you might use OpenCL to perform filter computations on the pixels of an image or use it to perform complex math calculations on several values at once. In other words, OpenCL is geared more toward problem sets whose data can be operated on in parallel.

Although OpenCL is good for performing massively data-parallel operations, it is not suitable for more general-purpose calculations. There is a nontrivial amount of effort required to prepare and transfer both the data and the required work kernel to a graphics card so that it can be operated on by a GPU. Similarly, there is a nontrivial amount of effort required to retrieve any results generated by OpenCL. As a result, any tasks that interact with the system are generally not recommended for use with OpenCL. For example, you would not use OpenCL to process data from files or network streams. Instead, the work you perform using OpenCL must be much more self-contained so that it can be transferred to the graphics processor and computed independently.

For more information about OpenCL and how you use it, see OpenCL Programming Guide for Mac OS X.

LemmycautioN · Aug 31, 2009, 05:01 AM

The first time i ran with the 9400 running and the 9600 wasn't recognized - also, notice the significant difference between the 9400 speeds

maviso5 · Aug 31, 2009, 06:56 AM

Code:

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce GT 120
Device 0 is an: GPU with max. 1400 MHz and 32 units/cores 
Now computing - please be patient....
time used:  1.658 seconds

OpenCL Device # 1 = Intel(R) Xeon(R) CPU           E5520  @ 2.27GHz
Device 1 is an: CPU with max. 2260 MHz and 16 units/cores 
Now computing - please be patient....
time used:  1.149 seconds

tacojohn · Aug 31, 2009, 12:32 PM

Quote:

Originally Posted by tacojohn

24" iMac 2.8GHz 4GB RAM 8800 GS 512VRAM

Code:

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce 8800 GS
Device 0 is an: GPU with max. 1250 MHz and 64 units/cores 
Now computing - please be patient....
time used:  0.924 seconds

OpenCL Device # 1 = Intel(R) Core(TM)2 Duo CPU     E8235  @ 2.80GHz
Device 1 is an: CPU with max. 2800 MHz and 2 units/cores 
Now computing - please be patient....
time used: 12.855 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

I thought my iMac would be slower than the dual GPU MBPs– I'd like to see some h.264 encoding results. I bet the 9400M would kick my iMac's ass...

nudlebaer · Sep 1, 2009, 04:23 AM

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 2
OpenCL Device # 0 = GeForce 8800 GT
Device 0 is an: GPU with max. 1500 MHz and 112 units/cores 
Now computing - please be patient....
time used:  0.692 seconds

OpenCL Device # 1 = Intel(R) Xeon(R) CPU           E5462  @ 2.80GHz
Device 1 is an: CPU with max. 2800 MHz and 8 units/cores 
Now computing - please be patient....
time used:  3.198 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

I can't believe Mac Pros are so fast compared to MacBooks

skm · Sep 1, 2009, 04:52 AM

When I'm running the 9400 I only see two devices. When I change to the 9600 I can see all 3...

Is this normal?

Also, these are my results, which look a bit odd...

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 3
OpenCL Device # 0 = GeForce 9600M GT
Device 0 is an: GPU with max. 1250 MHz and 32 units/cores 
Now computing - please be patient....
time used: 15.626 seconds

OpenCL Device # 1 = GeForce 9400M
Device 1 is an: GPU with max. 1100 MHz and 16 units/cores 
Now computing - please be patient....
time used:  6.835 seconds

OpenCL Device # 2 = Intel(R) Core(TM)2 Duo CPU     T9550  @ 2.66GHz
Device 2 is an: CPU with max. 2660 MHz and 2 units/cores 
Now computing - please be patient....
time used: 17.362 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

DAMNiatx · Sep 1, 2009, 10:43 AM

Quote:

Originally Posted by skm

When I'm running the 9400 I only see two devices. When I change to the 9600 I can see all 3...

Is this normal?

Also, these are my results, which look a bit odd...

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 3
OpenCL Device # 0 = GeForce 9600M GT
Device 0 is an: GPU with max. 1250 MHz and 32 units/cores 
Now computing - please be patient....
time used: 15.626 seconds

OpenCL Device # 1 = GeForce 9400M
Device 1 is an: GPU with max. 1100 MHz and 16 units/cores 
Now computing - please be patient....
time used:  6.835 seconds

OpenCL Device # 2 = Intel(R) Core(TM)2 Duo CPU     T9550  @ 2.66GHz
Device 2 is an: CPU with max. 2660 MHz and 2 units/cores 
Now computing - please be patient....
time used: 17.362 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

are you running it in battery ?

DAMNiatx · Sep 1, 2009, 12:07 PM

can we start with another benchmark ?,
because mitch_de make a comment about this benchmark

Quote:

My first o25 OpenCL Bench i, after longer tests, NOT USEFUL anymore. Not because of to less differences in the result, its the OpenCL work, which is to less real/to less complex (only an vector add).

new benchmark called qJulia OpenCL GPU FPS Bench
you can download it from this link
http://files.me.com/rheza/vupp26
This quote from insanelymac thread.

Quote:

qJulia OpenCL GPU FPS Bench
10.6 ONLY!
OPENCL - Good to know :
- OpenCL is an API for universal GPU(CPU) computing
- main difference to CUDA / ATI STEAM is: both only working with their "own" gpu.
an CUDA (NV) app like badaboom(h264 on GPU) cant work on an ATI gpu and vice versa
- OpenCL is universal for different gpu vendors means:
- Xcode / GCC compiles an code which includes the source (in C as an string) for the gpu programm
that c source is , different to CUDA/ATI STEAM , is compiled later by OpenCL Framework at runtime !
So same App can run on complete different gpus and also , without/less codechange om CPU if no
OpemCL gpu (newer ones) is found
The source (example below) for the gpu programm will be really compiled at runtime, not only interpreted.
So little differences between run of my bench may happen because of that compile on the run wink.gif

Does only GPU FPS, not CPU like the other benches here.
I am not sure if the Apple code can handle more than one GPU.
Source from Apple, minimal cosmetic changes (bigger window 800x800 vs 500x500 Apple = more GPU work)
compiled with LLVM GCC 4.2

Now, an really complex qJulia number qrunching on GPU with OpenGL view of the result & FPS.

Other than Galaxy or OpenCL Bench , it now has an external opencl source file (.cl).
That code (looks C code!) will be compiled by OpenCL framework at runtime and executed on GPU (really much gpu work!).
In OpenCL apps you can include such source (as source strings) or, as with this example use an external .cl.

My 9600GT gets around 29 FPS in the startsquence (changes each start a bit), with Animation (key SPACE)
from 16 - 60 FPS.
9600 GT / 29 FPS (16 - 60 FPS)

MacBook Pro 13", GPU GeForce 9400M
running at 6,25 fps (6-6,50)
heavy load OpenCL work is possible on such mobilesOpenCL gpus / low end OpenCL gpus but not usefull
Beside the gpu speed itself, such OpenCL parts need much own/real VRAM - if to less much VRAM-main memory swapping over PCIe

i'm asking mitch_de to make benchmark that run all device for do one calculation, so we can now the maximum points in one machine.

he said

Quote:

Not impossilbe but at this time OpenCL is very new and the Benches may vary much(good)or near no (bad) between different CPU/GPUs.

qJulia, in theory, should perform much more real world (number crunching, weather simulation, ....) OpenCL results.

If qJulia shows good differences between cards it can later be used for such an allinone Bench.
But first we must collect the results of very different gpus.

thanks guys, waiting your result, i will do benchmarks tonight

REBELinBLUE · Sep 1, 2009, 01:51 PM

Code:

...........................................................
.................. OpenCL Bench V 0.25 by mitch ...........
...... C2D 3GHz = 12 sec vs Nvidia 9600GT = 0,93 sec ......
... time results are not comparable to older version! .....
...........................................................

Number of OpenCL devices found: 3
OpenCL Device # 0 = GeForce 9600M GT
Device 0 is an: GPU with max. 1250 MHz and 32 units/cores 
Now computing - please be patient....
time used:  2.787 seconds

OpenCL Device # 1 = GeForce 9400M
Device 1 is an: GPU with max. 1100 MHz and 16 units/cores 
Now computing - please be patient....
time used: 15.235 seconds

OpenCL Device # 2 = Intel(R) Core(TM)2 Duo CPU     T9400  @ 2.53GHz
Device 2 is an: CPU with max. 2530 MHz and 2 units/cores 
Now computing - please be patient....
time used: 14.470 seconds

Now checking if results are valid - please be patient....
:) Validate test passed - GPU results=CPU results :)

Is it just me or is my GeForce 9400M slow :blink:

skm · Sep 1, 2009, 10:34 PM

Quote:

Originally Posted by DAMNiatx

are you running it in battery ?

Yeah I was - Does that make a difference?

I will run it plugged in again tonight.

Aug 30, 2009, 09:28 AM	#4
DAMNiatx macrumors 6502 Join Date: Jan 2009	hahahaha, no one care about it. i know why apple need update their product more often. __________________ MacRumors is a Forum Macrumors. We're like Yahoo! Answers. http://yit.me \| http://trendsbyminute.com

Aug 30, 2009, 11:13 AM	#5
jav6454 macrumors 601 Join Date: Nov 2007 Location: Ontario, CA \| Metairie, LA \| San Pedro Sula, Honduras	It incredible how the 9600M GT could process things in ~2 seconds where as the CPU took ~10 seconds. I know he difference o cores makes a difference. But dam! Never knew it'd make that much. __________________ Collector's Edition Al MacBook 2.4GHz \| 1st Gen iPhone 8GB \| iPhone 3G S⃣ 32GB New to iPhone World? Click Me On a search? MRoogle

Aug 31, 2009, 02:35 AM	#15
jpur macrumors member Join Date: Oct 2008	This were some of my runs from my June 2009 Macbook Pro. A couple of interesting things: 1. When I ran the test in 9400 mode, it did not recognize the 9600M 2. Running with 32 bit or 64 bit kernel made no difference 3. There was some variation in the 9400 benchmarks. Still is really cool to see the raw power that future SL programs will be able to tap into. Thanks for sharing J! Attached Thumbnails

Aug 31, 2009, 05:01 AM	#17
LemmycautioN macrumors member Join Date: Sep 2008 Location: Seattle	The first time i ran with the 9400 running and the 9600 wasn't recognized - also, notice the significant difference between the 9400 speeds __________________ 1st gen 15.4" Macbook Pro Unibody, 2.8 GHZ, 4 GB RAM, 250 GB 7200 RPM HDD - iPhone 8GB EDGE - iPod Classic 160GB

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Aug 31, 2009, 02:45 AM	#16
glycon macrumors newbie Join Date: Sep 2005	From Apple's page on Concurrency and Application Design (http://developer.apple.com/mac/libra...8091-CH100-SW1). The following Apple paragraphs seem a little contradictory, but this statement sums it up, "...any tasks that interact with the system are generally not recommended for use with OpenCL," in the second paragraph. (relevant portion from Apple page follows) OpenCL and Concurrency The Open Computing Language (OpenCL) is a standards-based technology for performing general-purpose computations on a computer’s graphics processor. OpenCL is a good technology to use if you have a well-defined set of computations that you want to apply to large data sets. For example, you might use OpenCL to perform filter computations on the pixels of an image or use it to perform complex math calculations on several values at once. In other words, OpenCL is geared more toward problem sets whose data can be operated on in parallel. Although OpenCL is good for performing massively data-parallel operations, it is not suitable for more general-purpose calculations. There is a nontrivial amount of effort required to prepare and transfer both the data and the required work kernel to a graphics card so that it can be operated on by a GPU. Similarly, there is a nontrivial amount of effort required to retrieve any results generated by OpenCL. As a result, any tasks that interact with the system are generally not recommended for use with OpenCL. For example, you would not use OpenCL to process data from files or network streams. Instead, the work you perform using OpenCL must be much more self-contained so that it can be transferred to the graphics processor and computed independently. For more information about OpenCL and how you use it, see OpenCL Programming Guide for Mac OS X.