<rellla>
one suspicious thing is, that PLBU_CMD_INDICES calculates the address, but info->start * info->index_size is not 0x40 aligned. not sure if this is one of the issues.
<rellla>
the logfile logs all activity on gp_submit and pp_submit and prints the bo tables after each lima_submit_add_bo and lima_submit_start
<rellla>
if there is any suspicious in refcnt or addresses, it should be visible now.
<rellla>
caching is disabled btw
yuq825 has quit [Quit: Leaving.]
maccraft123 has joined #lima
dddddd has joined #lima
monstr has quit [Remote host closed the connection]
chewitt has quit [Quit: Zzz..]
megi has quit [Ping timeout: 268 seconds]
<anarsoul>
rellla: will check it later today
megi has joined #lima
yann has quit [Ping timeout: 276 seconds]
megi has quit [Ping timeout: 268 seconds]
yann has joined #lima
maccraft123 has quit [Quit: WeeChat 2.6]
maccraft123 has joined #lima
piggz has joined #lima
maliciouscode1 has joined #lima
<maliciouscode1>
I do not know electronics that well, but I know mram technology being superior to transistor ones, because it scales to higher voltages via current mirrors, it is also nonvolatile and cheap and also lower-power hence, but should be generated via mixed analog digital toolkit.
<maliciouscode1>
So while usual digital bitline in-logic memristor circuits are for instance around with 14times lower switching speed capable, plugging a current mirror gates somewhere between a bitline memarray of that alu , it compares in performance the same or even better than transistor based logics
<maliciouscode1>
and have higher density and no wear out problems and all that stuff.
piggz has quit [Remote host closed the connection]
<maliciouscode1>
the fabrication of such alu arrays or manufacturing is cheap, and the machines to deposit materials are maskless and also cheap
<maliciouscode1>
tech is maskless I meant, so around 300k dollars you get a cvd machine working on very high 8000wph throughput
<maliciouscode1>
the beaming software for EBL can be optimized as well, to get higher throughput for that technology in addition to accelerating focusing with acoustics
<maliciouscode1>
but generally there is not much point in that, cause mram is just superior to anything really, the polarized magnet is the best tech imo.
<maliciouscode1>
there are lots of especially german and usa companies who are doing those sputtering and depositing machinery, the point was technology of that kind can be fabricated very quickly when status quo gets announced
<maliciouscode1>
There was a conadian thesis somewhere bookmarked which demonstrated any most popular alus underlying gates using mixed mode circuits with current mirrors added.
<maliciouscode1>
canadian
<maliciouscode1>
so the science behind appears to be done in that field as well.
maccraft123 has quit [Quit: WeeChat 2.6]
<maliciouscode1>
it is not close to shit from my fingers, this is something that already is known to have been worked out, in-logic memristor based alus
<maliciouscode1>
spintronics via magnetism, a electron tunneling method where voltage controls the polarism of the nanomagnet based gate
<maliciouscode1>
i.e memory cell
<maliciouscode1>
this takes no time when military equipment will be improved to really bigger heights than currently.