LuxMark v3.0: LuxRender propaganda with OpenCL

Discussions related to GPU Acceleration in LuxRender

Moderators: Dade, jromang, tomb, coordinators

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby lordchelo » Wed Apr 15, 2015 9:27 pm

The test "Neumann TLM-102 Special Edition (with EA-4 shock mount)" takes too long time to compile and the process uses a huge amount of RAM since I installed Nvidia drivers 349.12 (with support for OpenCL 1.2). The test system is Ubuntu 14.10 with a GeForce GTX 960.

Sorry for my bad English :)
lordchelo
 
Posts: 2
Joined: Wed Jun 09, 2010 5:28 pm

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby tomb » Thu Apr 16, 2015 1:54 am

Have you tried 350.12?
User avatar
tomb
Developer
 
Posts: 2677
Joined: Thu Oct 11, 2007 4:23 pm
Location: Oslo, Norway

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby lordchelo » Thu Apr 16, 2015 2:40 am

I have the same problem running the test from windows 7x64 with 350.12 drivers. With LuxBall HDR and Hotel lobby there are no problems.
lordchelo
 
Posts: 2
Joined: Wed Jun 09, 2010 5:28 pm

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby Dade » Thu Apr 16, 2015 2:56 am

lordchelo wrote:I have the same problem running the test from windows 7x64 with 350.12 drivers. With LuxBall HDR and Hotel lobby there are no problems.


It seems a problem (i.e. bug) related to latest drivers, 347.52 works fine.
User avatar
Dade
Developer
 
Posts: 8363
Joined: Sat Apr 19, 2008 6:04 pm
Location: Italy

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby richardlayman » Thu Apr 23, 2015 6:42 am

Is LuxMark 3.0 only for GPU devices - It used to work on my laptop but only on the HDR ball and the rest would crash with a seq fault.
With the current version I have (3.0-1) it doesn't even start.
Code: Select all
terminate called after throwing an instance of 'cl::Error'
  what():  clGetDeviceIDs
Aborted (core dumped)


Also, stuipid question, is the LuxMark score proportional the render times - say I have a machine that gets a score of 100 and another one that gets 1000, does that mean that the 1000 one would render the same scene x10 faster?
Feather
LuxEd - Lux Editor
ArchLinux 64bit
User avatar
richardlayman
 
Posts: 46
Joined: Sat Apr 18, 2015 7:58 pm

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby Dade » Thu Apr 23, 2015 8:18 am

richardlayman wrote:Is LuxMark 3.0 only for GPU devices - It used to work on my laptop but only on the HDR ball and the rest would crash with a seq fault.
With the current version I have (3.0-1) it doesn't even start.


What GPU do you have ? NVIDIA, AMD or Intel ?

For NVIDIA and AMD you need the very latest drivers. Intel GPUs are a bit hopeless for running large and resource intensive OpenCL applications like LuxRender/LuxMark.

richardlayman wrote:Also, stuipid question, is the LuxMark score proportional the render times - say I have a machine that gets a score of 100 and another one that gets 1000, does that mean that the 1000 one would render the same scene x10 faster?


The score is the number of samples per seconds: higher it is, more samples per second your hardware is capable to compute and less noise the rendered image has (in the same rendering time).

However, noise reduction is not linear (i.e. 10x faster => 10x less noise) but follow a square root (i.e. 10x faster => sqrt(10x) => 3.16x less noise). This is the one of the main reason why you need HUGE computing resources to render images.
User avatar
Dade
Developer
 
Posts: 8363
Joined: Sat Apr 19, 2008 6:04 pm
Location: Italy

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby richardlayman » Thu Apr 23, 2015 8:55 am

Thanks for the info Dade!

I'm using a Quad Core Intel
Code: Select all
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 58
model name      : Intel(R) Core(TM) i7-3537U CPU @ 2.00GHz
stepping        : 9
microcode       : 0x1b
cpu MHz         : 1584.863
cache size      : 4096 KB
physical id     : 0
siblings        : 4
core id         : 0
cpu cores       : 2
apicid          : 0
initial apicid  : 0
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm ida arat epb pln pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase smep erms xsaveopt
bugs            :
bogomips        : 4990.44
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management:

processor       : 1
vendor_id       : GenuineIntel
cpu family      : 6
model           : 58
model name      : Intel(R) Core(TM) i7-3537U CPU @ 2.00GHz
stepping        : 9
microcode       : 0x1b
cpu MHz         : 3085.937
cache size      : 4096 KB
physical id     : 0
siblings        : 4
core id         : 1
cpu cores       : 2
apicid          : 2
initial apicid  : 2
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm ida arat epb pln pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase smep erms xsaveopt
bugs            :
bogomips        : 4990.44
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management:

processor       : 2
vendor_id       : GenuineIntel
cpu family      : 6
model           : 58
model name      : Intel(R) Core(TM) i7-3537U CPU @ 2.00GHz
stepping        : 9
microcode       : 0x1b
cpu MHz         : 2899.902
cache size      : 4096 KB
physical id     : 0
siblings        : 4
core id         : 0
cpu cores       : 2
apicid          : 1
initial apicid  : 1
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm ida arat epb pln pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase smep erms xsaveopt
bugs            :
bogomips        : 4990.44
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management:

processor       : 3
vendor_id       : GenuineIntel
cpu family      : 6
model           : 58
model name      : Intel(R) Core(TM) i7-3537U CPU @ 2.00GHz
stepping        : 9
microcode       : 0x1b
cpu MHz         : 1586.230
cache size      : 4096 KB
physical id     : 0
siblings        : 4
core id         : 1
cpu cores       : 2
apicid          : 3
initial apicid  : 3
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm ida arat epb pln pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase smep erms xsaveopt
bugs            :
bogomips        : 4990.44
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management:


I'm using the i915 driver

Code: Select all
[work@laptop ~]$ pacman -Q xf86-video-intel
xf86-video-intel 2.99.917-5
[work@laptop ~]$ modinfo i915
filename:       /lib/modules/3.19.3-3-ARCH/kernel/drivers/gpu/drm/i915/i915.ko.gz
license:        GPL and additional rights
description:    Intel Graphics
author:         Intel Corporation
author:         Tungsten Graphics, Inc.
alias:          pci:v00008086d0000191Dsv*sd*bc03sc*i*
alias:          pci:v00008086d0000190Asv*sd*bc03sc*i*
alias:          pci:v00008086d0000192Asv*sd*bc03sc*i*
[I TOOK THESE alias's OUT JUST TO REDUCE THE OUTPUT]
...
depends:        drm_kms_helper,drm,intel-gtt,i2c-core,video,button,i2c-algo-bit
intree:         Y
vermagic:       3.19.3-3-ARCH SMP preempt mod_unload modversions
parm:           modeset:Use kernel modesetting [KMS] (0=DRM_I915_KMS from .config, 1=on, -1=force vga console preference [default]) (int)
parm:           panel_ignore_lid:Override lid status (0=autodetect, 1=autodetect disabled [default], -1=force lid closed, -2=force lid open) (int)
parm:           powersave:Enable powersavings, fbc, downclocking, etc. (default: true) (int)
parm:           semaphores:Use semaphores for inter-ring sync (default: -1 (use per-chip defaults)) (int)
parm:           enable_rc6:Enable power-saving render C-state 6. Different stages can be selected via bitmask values (0 = disable; 1 = enable rc6; 2 = enable deep rc6; 4 = enable deepest rc6). For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. default: -1 (use per-chip default) (int)
parm:           enable_fbc:Enable frame buffer compression for power savings (default: -1 (use per-chip default)) (int)
parm:           lvds_downclock:Use panel (LVDS/eDP) downclocking for power savings (default: false) (int)
parm:           lvds_channel_mode:Specify LVDS channel mode (0=probe BIOS [default], 1=single-channel, 2=dual-channel) (int)
parm:           lvds_use_ssc:Use Spread Spectrum Clock with panels [LVDS/eDP] (default: auto from VBT) (int)
parm:           vbt_sdvo_panel_type:Override/Ignore selection of SDVO panel mode in the VBT (-2=ignore, -1=auto [default], index in VBT BIOS table) (int)
parm:           reset:Attempt GPU resets (default: true) (bool)
parm:           enable_hangcheck:Periodically check GPU activity for detecting hangs. WARNING: Disabling this can cause system wide hangs. (default: true) (bool)
parm:           enable_ppgtt:Override PPGTT usage. (-1=auto [default], 0=disabled, 1=aliasing, 2=full) (int)
parm:           enable_execlists:Override execlists usage. (-1=auto, 0=disabled [default], 1=enabled) (int)
parm:           enable_psr:Enable PSR (default: false) (int)
parm:           preliminary_hw_support:Enable preliminary hardware support. (int)
parm:           disable_power_well:Disable the power well when possible (default: true) (int)
parm:           enable_ips:Enable IPS (default: true) (int)
parm:           fastboot:Try to skip unnecessary mode sets at boot time (default: false) (bool)
parm:           prefault_disable:Disable page prefaulting for pread/pwrite/reloc (default:false). For developers only. (bool)
parm:           invert_brightness:Invert backlight brightness (-1 force normal, 0 machine defaults, 1 force inversion), please report PCI device ID, subsystem vendor and subsystem device ID to dri-devel@lists.freedesktop.org, if your machine needs it. It will then be included in an upcoming module version. (int)
parm:           disable_display:Disable display (default: false) (bool)
parm:           disable_vtd_wa:Disable all VT-d workarounds (default: false) (bool)
parm:           enable_cmd_parser:Enable command parsing (1=enabled [default], 0=disabled) (int)
parm:           use_mmio_flip:use MMIO flips (-1=never, 0=driver discretion [default], 1=always) (int)
parm:           mmio_debug:Enable the MMIO debug code (default: false). This may negatively affect performance. (bool)
thor:         Intel Corporation
author:         Tungsten Graphics, Inc.
Feather
LuxEd - Lux Editor
ArchLinux 64bit
User avatar
richardlayman
 
Posts: 46
Joined: Sat Apr 18, 2015 7:58 pm

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby Dade » Thu Apr 23, 2015 10:32 am

richardlayman wrote:I'm using the i915 driver


Are you sure you weren't using OpenCL CPU device (i.e. not the GPU device) ? I don't think i915 GPU support OpenCL at all.
User avatar
Dade
Developer
 
Posts: 8363
Joined: Sat Apr 19, 2008 6:04 pm
Location: Italy

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby richardlayman » Thu Apr 23, 2015 12:14 pm

Sorry, forgot to mention that the intel-agp module gets loaded before the i915 - I believe I had to add it due to ACPI errors.
There are GPU options for the i915 but I'm not sure if this would be any relation to the issue I'm seeing:

Code: Select all
[work@laptop ~]$ modinfo -p i915
modeset:Use kernel modesetting [KMS] (0=DRM_I915_KMS from .config, 1=on, -1=force vga console preference [default]) (int)
panel_ignore_lid:Override lid status (0=autodetect, 1=autodetect disabled [default], -1=force lid closed, -2=force lid open) (int)
powersave:Enable powersavings, fbc, downclocking, etc. (default: true) (int)
semaphores:Use semaphores for inter-ring sync (default: -1 (use per-chip defaults)) (int)
enable_rc6:Enable power-saving render C-state 6. Different stages can be selected via bitmask values (0 = disable; 1 = enable rc6; 2 = enable deep rc6; 4 = enable deepest rc6). For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. default: -1 (use per-chip default) (int)
enable_fbc:Enable frame buffer compression for power savings (default: -1 (use per-chip default)) (int)
lvds_downclock:Use panel (LVDS/eDP) downclocking for power savings (default: false) (int)
lvds_channel_mode:Specify LVDS channel mode (0=probe BIOS [default], 1=single-channel, 2=dual-channel) (int)
lvds_use_ssc:Use Spread Spectrum Clock with panels [LVDS/eDP] (default: auto from VBT) (int)
vbt_sdvo_panel_type:Override/Ignore selection of SDVO panel mode in the VBT (-2=ignore, -1=auto [default], index in VBT BIOS table) (int)
reset:Attempt GPU resets (default: true) (bool)
enable_hangcheck:Periodically check GPU activity for detecting hangs. WARNING: Disabling this can cause system wide hangs. (default: true) (bool)
enable_ppgtt:Override PPGTT usage. (-1=auto [default], 0=disabled, 1=aliasing, 2=full) (int)
enable_execlists:Override execlists usage. (-1=auto, 0=disabled [default], 1=enabled) (int)
enable_psr:Enable PSR (default: false) (int)
preliminary_hw_support:Enable preliminary hardware support. (int)
disable_power_well:Disable the power well when possible (default: true) (int)
enable_ips:Enable IPS (default: true) (int)
fastboot:Try to skip unnecessary mode sets at boot time (default: false) (bool)
prefault_disable:Disable page prefaulting for pread/pwrite/reloc (default:false). For developers only. (bool)
invert_brightness:Invert backlight brightness (-1 force normal, 0 machine defaults, 1 force inversion), please report PCI device ID, subsystem vendor and subsystem device ID to dri-devel@lists.freedesktop.org, if your machine needs it. It will then be included in an upcoming module version. (int)
disable_display:Disable display (default: false) (bool)
disable_vtd_wa:Disable all VT-d workarounds (default: false) (bool)
enable_cmd_parser:Enable command parsing (1=enabled [default], 0=disabled) (int)
use_mmio_flip:use MMIO flips (-1=never, 0=driver discretion [default], 1=always) (int)
mmio_debug:Enable the MMIO debug code (default: false). This may negatively affect performance. (bool)


I have to make a correction on my previous post - the cpu is actually 2 cores with 4 threads.
This is my laptop so there are no GPU's, but that shouldn't make the program crash; it always worked in the past - you just couldn't render anything in GPU-only mode.
Feather
LuxEd - Lux Editor
ArchLinux 64bit
User avatar
richardlayman
 
Posts: 46
Joined: Sat Apr 18, 2015 7:58 pm

Re: LuxMark v3.0: LuxRender propaganda with OpenCL

Postby moby1kenobi » Tue Jun 30, 2015 7:33 am

Is the stress test any good for verifying an overclock? Has anyone used it for this?
User avatar
moby1kenobi
 
Posts: 41
Joined: Sun Jun 16, 2013 7:39 am

PreviousNext

Return to GPU Acceleration

Who is online

Users browsing this forum: Exabot [Bot] and 1 guest