Con Kolivas
c1ff90a7ab
Do testing for HW errors on submit nonce for both scrypt and sha.
12 years ago
Con Kolivas
6b714dd431
Increment hardware error count from the one site.
12 years ago
Con Kolivas
5412323e26
Fix memory leak with share submission on GPU work structures as discovered by twobitcoins.
12 years ago
Kano
a344deb6ac
rename device_api -> device_drv and all related api -> drv and add a device_drv->drv enum for identifying which driver each is
12 years ago
Con Kolivas
c99636f21b
Begin tearing down the old workio command queues by removing submit commands from there and submit them asynchronously via their own threads.
12 years ago
Con Kolivas
8bd8696854
The work struct pointer in struct pc_data in findnonce is never freed yet there is no need to allocate it separately so make struct work a static part of the struct pc_data.
...
s
12 years ago
Con Kolivas
2b6e3676cd
Count an invalid nonce count as a hardware error on opencl.
12 years ago
Con Kolivas
2ff1610307
Prevent corrupt values returned from the opencl code from trying to read beyond the end of the buffer by masking the value to a max of 15.
12 years ago
Kano
307d8da034
HW: error counter auto for all devices - ztex code not fixed
12 years ago
Con Kolivas
86d5377e72
Use atomic ops to never miss a nonce on opencl kernels, including nonce==0, also allowing us to make the output buffer smaller.
13 years ago
Con Kolivas
e49bd98196
Use the scrypt CPU code to confirm results from OCL code, and mark failures as HW errors, making it easier to tune scrypt parameters.
13 years ago
Con Kolivas
b9e5f8e550
Revert "Use cpu scrypt code to check if an scrypt share is below target before submitting it."
...
This reverts commit 808f403a14
.
Conflicts:
findnonce.c
13 years ago
Con Kolivas
d9eba43a42
Fix build.
13 years ago
Con Kolivas
808f403a14
Use cpu scrypt code to check if an scrypt share is below target before submitting it.
13 years ago
Con Kolivas
d72add9af3
Send correct values to scrypt kernel to get it finally working.
13 years ago
Con Kolivas
6ac14f4280
Don't enqueuewrite buffer at all for pad8 and pass work details around for scrypt in dev_blk.
13 years ago
Con Kolivas
ea444d0239
Fix nonce submission code for scrypt.
13 years ago
Con Kolivas
8fd3bf74df
Build fix for opt scrypt.
13 years ago
Con Kolivas
1aacfe5279
Don't check postcalc nonce with sha256 in scrypt.
13 years ago
Con Kolivas
b085c338f6
Make scrypt buffers and midstate compatible with cgminer.
13 years ago
Con Kolivas
8f08a775ad
Use any() in kernel output code and revert breakage of diakgcn kernel.
13 years ago
Con Kolivas
d1cddf8bad
Update licensing to GPL V3.
13 years ago
Con Kolivas
810ad04578
More copyright updates.
13 years ago
Con Kolivas
83dde50f9d
Updates to diakgcn kernel courtesy of Philip Kaufmann <phil.kaufmann@t-online.de>
13 years ago
zefir
713e8be629
move CPU chipset specific optimization into device-cpu
13 years ago
Con Kolivas
52bb99c3ea
Fix double definition of A0 B0 to zeroA zeroB.
13 years ago
Con Kolivas
405a2120f8
Remove unnecessary check for opt_debug on every invocation of applog at LOG_DEBUG and place the check in applog().
13 years ago
Philip Kaufmann
c91155ac55
made code a bit more consistent and removed 2 unneeded brackets
13 years ago
ckolivas
85e9a841d5
Use K array explicitly to make it clear what is being added.
13 years ago
ckolivas
033913ca8e
First working port of the diakgcn kernel.
13 years ago
ckolivas
1bfa82c3ce
Revert "Search for nonce if it is not found to be correct as a debugging check."
...
This reverts commit d9e07385da
.
Purely a debugging exercise.
13 years ago
ckolivas
d9e07385da
Search for nonce if it is not found to be correct as a debugging check.
13 years ago
Philip Kaufmann
68e36c6573
made code a bit more consistent and removed 2 unneeded brackets
13 years ago
ckolivas
7a069b93af
Use K array explicitly to make it clear what is being added.
13 years ago
ckolivas
53d33c753f
First working port of the diakgcn kernel.
13 years ago
Con Kolivas
8171ceab85
Simplify submit_nonce loop and avoid potentially missing FOUND - 1 entry. Reported by Luke-Jr.
13 years ago
Con Kolivas
a51514d9d1
White space cleanup.
13 years ago
Con Kolivas
4d13f8ad3c
ByteReverse is not used and the bswap opcode breaks big endian builds. Remove it.
13 years ago
Con Kolivas
cf54f9b850
Move to 256 sized buffers and don't risk overwrite by using only 127 mask.
14 years ago
Con Kolivas
0f782ba6bd
Update poclbm kernel to FF sized mask and only check that range.
14 years ago
Phateus
d15d225a4c
Changed phatk version to 2.2
14 years ago
Con Kolivas
116a9dc025
Update phatk kernel to one with new parameters for slightly less overhead again.
...
Make the queue kernel parameters call a function pointer to select phatk or poclbm.
14 years ago
Con Kolivas
1c070475c3
Cycling over 1026 nonces is a remnant of old code not relevant now and can just lead to repeat submission of the same block increasing rejects.
...
Remove it.
14 years ago
Con Kolivas
b6ac9f185e
Revert "Make sure not to try and submit work for nonces higher than already submitted in current work item."
...
This reverts commit a2dd410e36
.
Incorrect fix. Missed real shares that are accepted.
14 years ago
Con Kolivas
a2dd410e36
Make sure not to try and submit work for nonces higher than already submitted in current work item.
...
This reduces rejects substantially.
14 years ago
Con Kolivas
cb13e2cfe5
Make it possible to build without opencl for cpu mining only.
14 years ago
Con Kolivas
13b43cfad1
Update copyright and authors.
14 years ago
ckolivas
7ae9afc40f
Profile points and warning clean ups.
14 years ago
Con Kolivas
d5d4d1da16
Don't want to free the work data out of the transient structs.
14 years ago
Con Kolivas
2b6e841673
Use a buffer of up to 512 * 4 integers when retrieving work from the GPU.
...
This allows each local thread id to have one slot to put any positive results into, thus making overlapping results far less likely.
Thus races will be much rarer, allowing more threads.
It should also pick up blocks close to each other more reliably and hopefully decrease the number of rejects and opencl errors.
Do the search over the buffer entirely in a separate thread to allow the GPU to stay as busy as possible.
Detach threads from themselves to prevent unlucky even where dereferencing occurs by freeing the data that stores the thread info.
14 years ago