Go to advanced search

by hglm
Sun Mar 15, 2015 2:43 pm
Forum: General discussion
Topic: Faulty Pi Memory
Replies: 45
Views: 14489

Re: Faulty Pi Memory

The Pi2 generally cannot be overclocked as fast as the Pi1, for various reasons. Which I cannot remember although had been told at some point in the past. The Raspberry Pi 2 processor chip uses entirely different ARM processor cores and an entirely different type of RAM. Considering this significan...
by hglm
Tue Mar 10, 2015 3:11 pm
Forum: Advanced users
Topic: Overclocking
Replies: 1125
Views: 605454

Re: Overclocking

I found your write-up very helpful and have had good success with the 1100+4/550/483 setting shown at the top of your table of stable high-performance clock configurations. A few questions: I noticed your settings seem to go up in 16.667MHz steps (450/467/483/500). Is there any particular reason fo...
by hglm
Sun Mar 08, 2015 1:26 pm
Forum: General discussion
Topic: SD Card Benchmarks
Replies: 301
Views: 295567

Re: SD Card Benchmarks

I'm using the flash-bench results (run on the RPi 2 itself) to compare different mSD Cards benchmarks, as I can see little correlation with the results produced on a Windows computer with CrystalDiskMark? The Raspberry Pi 1 and 2 have a slower SD-card interface than most PCs, partly because it uses...
by hglm
Tue Feb 24, 2015 4:14 pm
Forum: General discussion
Topic: SD Card Benchmarks
Replies: 301
Views: 295567

Re: SD Card Benchmarks

Using your benchmark, the results for my Samsung 64GB Evo (MB-MP64D) (my previous iozone3 benchmarks above were also recorded on the Pi2) Seq. Seq. Random Random Read Write Read Write Raspberry Pi 2 17.7 13.9 5.49 3.65 That establishes that Samsung card as a good performer on the RPi 2, with random...
by hglm
Tue Feb 24, 2015 10:10 am
Forum: General discussion
Topic: SD Card Benchmarks
Replies: 301
Views: 295567

Re: SD Card Benchmarks

It would make sense to use a "standard" sd test method run on the Raspberry Pi itself. I see a wide variety of benchmarks on here and not many look to have been generated on the Raspberry Pi. Is there a "standard" tool that we could use to create a database of real Pi statistics? I am sure there ar...
by hglm
Tue Feb 24, 2015 12:45 am
Forum: Graphics programming
Topic: Low-level (Linux framebuffer) graphics programming tutorial
Replies: 62
Views: 77759

Re: Low-level (Linux framebuffer) graphics programming tutor

You might be interested in this: I experimented with the kerrnel fb layer driver for the Raspberry Pi 1/2 and it turns out it is pretty straightforward to extend the framebuffer visible to console applications from one screen buffer to three buffers. This makes double or triple-buffering possible fo...
by hglm
Fri Feb 20, 2015 4:18 pm
Forum: Advanced users
Topic: Overclocking
Replies: 1125
Views: 605454

Re: Overclocking

With the Raspberry Pi 2 Model B, it doesn't seem like I'm able to affect/tweak the GPU (h264, isp, v3d) clocks directly any longer or maybe the relationships of some of the clocks no longer behaves as on the B+. No matter what I do, I can't get seem to get those clocks over 250000000. The "core_fre...
by hglm
Mon Feb 16, 2015 12:41 pm
Forum: OpenGLES
Topic: GLFW3 initialisation
Replies: 22
Views: 7730

Console GLES2 apps in an X window and supporting X11 GLES2 a

Does that work for you? github.com/kika123/x11eglrpi That uses a hack that reads pixels from gpu into cpu then writes them back to the Xserver. Not a real solution in my opinion. PeterO It is not slow in practice as the copy uses DMA. I think such a solution might be workable but I don't think the ...
by hglm
Mon Feb 16, 2015 12:13 pm
Forum: Graphics programming
Topic: hello_videocube video is not smooth
Replies: 3
Views: 1648

Re: hello_videocube video is not smooth

It is the nature of a 3D GPU that coordinate transformations like movement and rotations do not really add to the amount of calculations that are performed, so speed is relatively unaffected. The GPU has to draw the video cube every frame with a similar amount of calculations even when the cube is s...
by hglm
Tue Jul 02, 2013 11:08 am
Forum: Linux Kernel
Topic: Building a new kernel
Replies: 4
Views: 10196

Re: Building a new kernel

Does this mean I should just do: sudo cp ~/modules / ...and copy my just made modules right over the top of the exisiting installed modules - this seems like it could break something, and also gives me no software backout option to step back to the eixisting build of modules? Or am I just over thin...
by hglm
Wed Jun 26, 2013 10:18 pm
Forum: Linux Kernel
Topic: Optimizing the kernel copy_page and memcpy functions
Replies: 9
Views: 7898

Re: Optimizing the kernel copy_page and memcpy functions

I have been doing some more work on this and I think am making some progress. The validation errors I was seeing with the original kernel memcpy have cleared up (accidently set wrong endianness in my userspace environment). The memcpy implementation now performs more highly tuned preloads and has a ...
by hglm
Sat Jun 22, 2013 7:52 pm
Forum: Linux Kernel
Topic: Optimizing the kernel copy_page and memcpy functions
Replies: 9
Views: 7898

Re: Optimizing the kernel copy_page and memcpy functions

In the latest version of the patch I fixed a performance regression in memset/memzero which was caused by forcing write alignment to a 32-byte boundary, which was beneficial for the memcpy code. The RPi doesn't like memset/memzero writing whole cache lines at once at an aligned address. Profling "pe...
by hglm
Sat Jun 22, 2013 2:54 pm
Forum: Linux Kernel
Topic: Optimizing the kernel copy_page and memcpy functions
Replies: 9
Views: 7898

Re: Optimizing the kernel copy_page and memcpy functions

How does your memcpy compare with the ones we use in userland (https://github.com/bavison/arm-mem/)? How about memset? I believe Raspbian uses the libcofi optimized memcpy and memset via the ld.so.preload mechanism (/etc/ld.so.preload), which is from https://github.com/simonjhall/copies-and-fills/,...
by hglm
Sat Jun 22, 2013 2:19 pm
Forum: General discussion
Topic: Experimental enhanced X driver (rpifb)
Replies: 98
Views: 54374

Re: Experimental enhanced X driver (rpifb)

DMA can be used for fast filling. Can this be plumbed into X? I implemented accelerated filling in the sunxifb driver on the sunxi platform (which uses a 2D DMA engine analogous to the RPi), and the consensus was that this is not a really big win, except in some contrived cases. because direct fram...
by hglm
Sat Jun 22, 2013 12:45 pm
Forum: Linux Kernel
Topic: Optimizing the kernel copy_page and memcpy functions
Replies: 9
Views: 7898

Re: Optimizing the kernel copy_page and memcpy functions

I did attempt this a while back, but found running the quake3 timedemo would sometimes segfault with the patch applied. I had a suspicion it was misaligned access related, but haven't had a chance to investigate. The copy_page patch is quite straightforward (since copy_page is always page aligned),...
by hglm
Sat Jun 22, 2013 10:18 am
Forum: Linux Kernel
Topic: Optimizing the kernel copy_page and memcpy functions
Replies: 9
Views: 7898

Re: Optimizing the kernel copy_page and memcpy functions

I have benchmarked a kernel with optimized copy_page and memcpy functions. To do so, I had to compile OProfile from source and make sure it uses the timer-based sampling method instead of hardware performance counters which are not fully supported by the ARMv6 platform. Using oprofile, and supplying...
by hglm
Sat Jun 22, 2013 10:07 am
Forum: Linux Kernel
Topic: Optimizing the kernel copy_page and memcpy functions
Replies: 9
Views: 7898

Optimizing the kernel copy_page and memcpy functions

While memcpy in userspace has received plenty of attention to be optimized for the Raspberry Pi, the same cannot be said for the memcpy-related functions in the kernel, the performance of which can be important for certain workloads. The RPi is very sensitive to the right prefetch strategy when copy...
by hglm
Fri Jun 21, 2013 8:53 pm
Forum: General discussion
Topic: R on Raspberry Pi
Replies: 4
Views: 883

Re: R on Raspberry Pi

Code: Select all

sudo apt-cache search r-cran
returns a long list of R-related packages available on Raspbian.
by hglm
Wed Jun 19, 2013 4:53 pm
Forum: General discussion
Topic: Experimental enhanced X driver (rpifb)
Replies: 98
Views: 54374

Re: Experimental enhanced X driver (rpifb)

Unfortunately it looks heavy file system access concurrent with DMA scrolling can cause a hard crash. I tried the following program to concurrently: (1) Cat a big text file to the console, triggering almost continuous DMA scrolling. (2) Read a big file from the root file system (SD card). The result...
by hglm
Tue Jun 18, 2013 7:59 pm
Forum: General discussion
Topic: Experimental enhanced X driver (rpifb)
Replies: 98
Views: 54374

Re: Experimental enhanced X driver (rpifb)

That's a good point, but I don't think it directly applies here. The CPU imageblit implementation is unsurprisingly fully occupying the CPU while it is doing its stuff, so the concurrently running programs are totally out of luck and are just waiting to be scheduled. Even if less memory bandwidth i...
by hglm
Tue Jun 18, 2013 12:47 am
Forum: General discussion
Topic: Experimental enhanced X driver (rpifb)
Replies: 98
Views: 54374

Re: Experimental enhanced X driver (rpifb)

That's a good point, but I don't think it directly applies here. The CPU imageblit implementation is unsurprisingly fully occupying the CPU while it is doing its stuff, so the concurrently running programs are totally out of luck and are just waiting to be scheduled. Even if less memory bandwidth i...
by hglm
Mon Jun 17, 2013 10:29 pm
Forum: General discussion
Topic: Experimental enhanced X driver (rpifb)
Replies: 98
Views: 54374

Re: Experimental enhanced X driver (rpifb)

There is only a minimal performance improvement for framebuffer console scrolling (just something like ~20% faster for 1280x720 at 32bpp and only marginally faster for 16bpp). My understanding is that fbcon tries to avoid reading back from the framebuffer by default and instead draws text at the ne...
by hglm
Mon Jun 17, 2013 8:29 pm
Forum: General programming discussion
Topic: Speeding up the nano editor
Replies: 1
Views: 1120

Speeding up the nano editor

nano is quite a useful editor, especially in the console, useful for system administration tasks. When using it for programming, I was frustrated by lag and slowness for moderately sized files and larger ones. I was putting this down to the slowness of the Pi, but it turns out it has little to do w...
by hglm
Mon Jun 17, 2013 5:10 pm
Forum: General discussion
Topic: Experimental enhanced X driver (rpifb)
Replies: 98
Views: 54374

Re: Experimental enhanced X driver (rpifb)

That's impressive. Its looks like a nice step forwards for the RPi user experience. The bonus of accelerated console scrolling is very welcome too. Only drawback is it requires a kernel patch, but I guess it should be possible to get it integrated into the standard Raspbian kernel within reasonable ...
by hglm
Sat Jun 15, 2013 11:02 pm
Forum: General discussion
Topic: Experimental enhanced X driver (rpifb)
Replies: 98
Views: 54374

Re: Experimental enhanced X driver (rpifb)

I do realize that the GPU is the most powerful feature on the RPi and mostly wasted at the moment, and that X is not the way forward with regard to fully utilizing it. My comment was aimed at the general notion that the slowness of X on the Pi is completely unaddressable and that Wayland would be th...

Go to advanced search