PTX - revision history

Dank at 09:23, 14 December 2010

2010-12-14T09:23:11Z

← Older revision		Revision as of 09:23, 14 December 2010
Line 28:		Line 28:
	\| 3.0		\| 3.0
	\|		\|
			* Added special registers <tt>nsmid</tt>, <tt>lanemask_*</tt>, <tt>clock64</tt>
			\|-
			\| 1.3
			\|
			\|
			* Added special registers <tt>laneid</tt>, <tt>warpid</tt>, <tt>smid</tt>, <tt>pm0-3</tt>
			* Added instructions <tt>sub.cc</tt> and <tt>sub.c</tt>
	\|-		\|-
	\|}		\|}

Dank at 08:40, 14 December 2010

2010-12-14T08:40:38Z

← Older revision		Revision as of 08:40, 14 December 2010
Line 1:		Line 1:
	The ISA to which [[CUDA]]'s nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a PTX module. It can then be scheduled for execution on CUDA devices. From Version 2.1 of the [http://developer.download.nvidia.com/compute/cuda/3_1/toolkit/docs/ptx_isa_2.1.pdf PTX ISA Reference]:		The ISA to which [[CUDA]]'s nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a PTX module. It can then be scheduled for execution on CUDA devices. From Version 2.1 of the [http://developer.download.nvidia.com/compute/cuda/3_1/toolkit/docs/ptx_isa_2.1.pdf PTX ISA Reference]:
	:''PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.''		:''PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.''

			All PTX instructions can be predicated. Predicated instances of the branch and call instructions are the only way to effect conditional branching.
	==Versions==		==Versions==
	{\| border="1"		{\| border="1"

Dank at 08:06, 14 December 2010

2010-12-14T08:06:29Z

← Older revision		Revision as of 08:06, 14 December 2010
Line 28:		Line 28:
	\|-		\|-
	\|}		\|}
			==Cooperative Thread Arrays==
			* Equivalent to a ''block'' in [[CUDA]] -- broken up into warps, can communicate, can be grouped into a grid, one kernel per grid
			* <tt>tid</tt>: thread ID within CTA
			* <tt>ntid</tt>: 3D shape of CTA
			* <tt>ctaid</tt>: CTA ID within grid
			* <tt>nctaid</tt>: 3D shape of grid
			* <tt>gridid</tt>: grid ID
	==Tools==		==Tools==
	* Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler		* Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler
	[[CATEGORY:GPGPU]]		[[CATEGORY:GPGPU]]

Dank: add PTX 2.2 notes

2010-12-14T07:52:56Z

add PTX 2.2 notes

← Older revision		Revision as of 07:52, 14 December 2010
Line 7:		Line 7:
	! CUDA Toolkit Version		! CUDA Toolkit Version
	! Changes		! Changes
			\|-
			\| 2.2
			\| 3.2
			\|
			* New kernel parameter directives for pointer arguments
			* Flat address space for constants (backwards compatibility for constant banks)
			* Texture changes for OpenCL, bilerp (bilinear interpolation) and high-bw loads
	\|-		\|-
	\| 2.1		\| 2.1

Dank at 07:42, 18 July 2010

2010-07-18T07:42:08Z

← Older revision		Revision as of 07:42, 18 July 2010
Line 2:		Line 2:
	:''PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.''		:''PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.''
	==Versions==		==Versions==
	{\|		{\| border="1"
	\|-		\|-
	! PTX Version		! PTX Version

Dank at 07:41, 18 July 2010

2010-07-18T07:41:58Z

@@ Line 1: / Line 1: @@
 The ISA to which [[CUDA]]'s nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a PTX module. It can then be scheduled for execution on CUDA devices. From Version 2.1 of the [http://developer.download.nvidia.com/compute/cuda/3_1/toolkit/docs/ptx_isa_2.1.pdf PTX ISA Reference]:
 :''PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.''
 ==Tools==
 * Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler
 [[CATEGORY:GPGPU]]

Dank at 07:33, 18 July 2010

2010-07-18T07:33:16Z

← Older revision		Revision as of 07:33, 18 July 2010
Line 3:		Line 3:
	==Tools==		==Tools==
	* Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler		* Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler
			[[CATEGORY:GPGPU]]

Dank at 07:33, 18 July 2010

2010-07-18T07:33:02Z

← Older revision		Revision as of 07:33, 18 July 2010
Line 1:		Line 1:
	The ISA to which [[CUDA]]'s nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a PTX module. It can then be scheduled for execution on CUDA devices. From Version 2.1 of the [http://developer.download.nvidia.com/compute/cuda/3_1/toolkit/docs/ptx_isa_2.1.pdf PTX ISA Reference]:		The ISA to which [[CUDA]]'s nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a PTX module. It can then be scheduled for execution on CUDA devices. From Version 2.1 of the [http://developer.download.nvidia.com/compute/cuda/3_1/toolkit/docs/ptx_isa_2.1.pdf PTX ISA Reference]:
	:PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.		:''PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.''
	==Tools==		==Tools==
	* Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler		* Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler

Dank: Created page with 'The ISA to which CUDA's nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a P...'

2010-07-18T07:32:46Z

Created page with 'The ISA to which CUDA's nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a P...'

New page

The ISA to which [[CUDA]]'s nvcc compiles source code. This is JIT'd into architecture-specific machine language by the hardware driver after the CUDA runtime is used to load a PTX module. It can then be scheduled for execution on CUDA devices. From Version 2.1 of the [http://developer.download.nvidia.com/compute/cuda/3_1/toolkit/docs/ptx_isa_2.1.pdf PTX ISA Reference]:
:PTX defines a virtual machine and ISA for general purpose parallel thread execution. PTX programs are translated at install time to the target hardware instruction set. The PTX-to-GPU translator and driver enable NVIDIA GPUs to be used as programmable parallel computers.
==Tools==
* Marcin Wilhelm Kościelnicki's [http://0x04.net/cgit/index.cgi/nv50dis/ nv50dis], a disassembler