Update Readme: CUDA, CMake #25
Conversation
|
|
||
| #if you are using AMD GPUs, uncomment the following line and set the install path correctly | ||
| #TARGET=ocl_memtest | ||
| AMD_INSTALL_PATH ?=/usr/local/ati-stream-sdk-v2.1-lnx64/ |
There was a problem hiding this comment.
wow, I am old enough to remember ATI stream.
There was a problem hiding this comment.
I did use it in the good old times xD
Update the readme with CMake notes, CUDA-focus of the fork, newer instructions and authors.
ax3l
force-pushed
the
doc-updateReadme
branch
8 times, most recently
from
3859027 to
64f3c0d
Compare
ax3l
force-pushed
the
doc-updateReadme
branch
from
64f3c0d to
88a975f
Compare
|
|
||
| ### Compile | ||
|
|
||
| Inside the source directory, run: |
There was a problem hiding this comment.
Don't we normally recommend an out-of-source build? Guess for this in-source should be fine when done like described here.
There was a problem hiding this comment.
Running in a quick-&-dirty but empty dir is not perfect but good enough. Simplifies instructions below.
| the address wires. | ||
|
|
||
| Test 1 `[Own address test]` | ||
| Each Memory location is filled with its own address. The next kernel checks if the |
|
|
||
| ### Known Issues | ||
|
|
||
| * If your machine is cuda 2.2, killing the program while it is running test 10 (the memory stress test) could result |
There was a problem hiding this comment.
machine is cuda 2.2 should perhaps be machine runs cuda 2.2 ?
| ### Known Issues | ||
|
|
||
| * If your machine is cuda 2.2, killing the program while it is running test 10 (the memory stress test) could result | ||
| in your GPUs in bad state. This is a bug from the nvidia driver. A detailed description can be found in |
There was a problem hiding this comment.
I think there is a word missing between GPUs and in, like being or remaining.
There was a problem hiding this comment.
nvidia is also written as Nvidia, both versions are in several places.
There was a problem hiding this comment.
Or retro NVidia or nVidia? Nobody knows
| Then we exit the kernel so that the memory can be flushed. Then we start a new kernel to read | ||
| and check if the value matches the pattern. An error is recorded if it does not match for each | ||
| memory location. In the same kernel, the compliment of the pattern is written after the checking. | ||
| The third kernel is launched to read the value again and checks against the compliment of the pattern. |
| ### Detailed Description | ||
|
|
||
| Test 0 `[Walking 1 bit]` | ||
| This test changes one bit a time in memory address to see it |
| are completed the data patterns are checked. Because the data is checked | ||
| only after the memory moves are completed it is not possible to know | ||
| where the error occurred. The addresses reported are only for where the | ||
| bad pattern was found. |
There was a problem hiding this comment.
This paragraph has double spaces between sentences.
|
|
||
| Test 8 `[Modulo 20, random pattern]` | ||
| A random pattern is generated. This pattern is used to set every 20th memory location | ||
| in memory. The rest of the memory location is set to the complimemnt of the pattern. |
There was a problem hiding this comment.
in memory is excessive, due to memory location.
There was a problem hiding this comment.
The rest of the memory location is - shouldn't it be plural?
| The bit fade test initializes all of memory with a pattern and then | ||
| sleeps for 90 minutes. Then memory is examined to see if any memory bits | ||
| have changed. All ones and all zero patterns are used. This test takes | ||
| 3 hours to complete. The Bit Fade test is disabled by default |
There was a problem hiding this comment.
Here is says 3 hours, but in the name 90 minutes.
| Test 10 `[memory stress test]` | ||
| Stress memory as much as we can. A random pattern is generated and a kernel of large grid size | ||
| and block size is launched to set all memory to the pattern. A new read and write kernel is launched | ||
| immediately after the previous write kernel to check if there is any errors in memory and set the |
There was a problem hiding this comment.
there is any errors is singular and plural mismatch.
| immediately after the previous write kernel to check if there is any errors in memory and set the | ||
| memory to the compliment. This process is repeated for 1000 times for one pattern. The kernel is | ||
| written as to achieve the maximum bandwidth between the global memory and GPU. | ||
| This will increase the chance of catching software error. In practice, we found this test quite useful |
There was a problem hiding this comment.
software error should either be plural or singular with a in front.
|
I did not change the original lingo, but you are welcome to just add a follow-up PR ;) |
Update the readme with CMake notes, CUDA-focus of the fork, newer instructions and authors.
Remove old
Makefile, we do not use nor support that.Related to #22 #24