|
Con,
What about Avalon cutoff temp? Will you do it?
|
|
|
|
So I thought I'd make it easy for everyone to try the latest cgminer 3.3.0 and overclock settings all rolled into one, so here is a firmware with the overclock settings settable from the web interface:
 (perhaps my mistake was that I have kept settings) I got this same exact Cgminer API error HELP ME PLEASE! My unit has been offline for 8 hours now  kill cgminer process (needs to be done twice) save config again.... try 345 instead of 350... it seems more stable for me goto startup and RESTART cgminer..... repeat this process every 5mins if not working Apiallow have you touched it ? |
|
|
|
NEWS FLASH!!! NEWS FLASH!!!
My K16 is hashing with 1 chip.
It's not giving a correct result nonce yet but I'm almost 99% sure that's due to errors in precalc code or maybe I'm shifting in word order backwards.
I soldered an ASIC on the board this morning and spent all day twiddling this and that, mostly getting the clock config right. Then bamm I saw something that looked like nonce bits coming out but they were stunted little bits trying desperately to reach high. But the twinkle was enough and I went through the Avalon reference design again to check and realized I'd not used the correct pull-up value for the result lines. Somehow I f'd up and spec'd 100k resistors instead of 470R. So I quickly removed them and soldered in some 1K that I had here. And the bits sprung forth to their full heights. I was totally excited and hopping around my work bench. Even my wife was in to see what the big commotion was about.
So... next step, test the UART receiver, and muck with the send data until the right nonce comes back.
BTW I'm running at half clock b/c no heat sink attached yet. The chip gets slightly warmish to the finger. With only one chip and half clock you have to set the scope trigger and hold the probes for a while until a nonce is found. I counted about 4-8 seconds but of course it depends on what point it is in it's cycle. One chip at half clock (128MHz) would be about 33 seconds for a full sweep.
GEAT! We are moving forward  |
|
|
|
So I thought I'd make it easy for everyone to try the latest cgminer 3.3.0 and overclock settings all rolled into one, so here is a firmware with the overclock settings settable from the web interface: http://ck.kolivas.org/apps/cgminer/avalon/20130624/openwrt-ar71xx-generic-tl-wr703n-v1-squashfs-factory.binNote that I have donation settings in the configuration so if you do not use your own config settings, it will start mining for me until you change them. Usual warnings apply with aggressive overclocking: do so at your own risk! Enjoy Thank you ckolivas! If you could add the 5Mhz steps from strombom into your firmware and the ability to change it via web interface, I will send you a tip of 0.5BTC I am inspecting the info provided from strombom There are difference between registers Form latest con cgminer source switch (frequency) { case 256: buf[6] = 0x03; buf[7] = 0x08; break; default: case 270: buf[6] = 0x73; buf[7] = 0x08; break; case 282: buf[6] = 0xd3; buf[7] = 0x08; break; case 300: buf[6] = 0x63; buf[7] = 0x09; break; case 325: buf[6] = 0x28; buf[7] = 0x0a; break; case 350: buf[6] = 0xf0; buf[7] = 0x0a; break; case 375: buf[6] = 0xb8; buf[7] = 0x0b; break; } Form hacked cgminer 5 Mhz steps } else if (frequency == 325) { buf[6] = 0x2b; buf[7] = 0x0a; } else if (frequency == 350) { buf[6] = 0xf3; buf[7] = 0x0a; } else if (frequency == 375) { buf[6] = 0xbb; buf[7] = 0x0b; I will not patch mine cgminer unless con decides to implement it Best PS: Att to strombom Can you comment why values are different and timing also? Thank you! Best |
|
|
|
Thank You! I was the one to ask best |
|
|
|
I've added custom temperature control to the avalon code in the cgminer git master.
It now acts like a simple PID controller trying to maintain a target temperature, which by default is set to 45 degrees, so it will try to keep the temp 43~45, but never going below a minimum fanspeed of 20% as a safety precaution and to maintain airflow.
You can modify this value with the new command:
--avalon-temp
Setting this to something very low will have the same effect as turning the fans up to maximum and vice versa.
If you are not running it from a command line, it can be added to the "More options" box in the web interface (once you have the binary on your machine).
Con, --avalon-temp is a great feature. Can you be so kind and implement --avalon-cutoff-temp - Gpu style to put avalon in IDLE. I personally run my unit at home with air condition. Outside temp is 40+ Celsius. And i am away most of the time. So if the air conditioner malfunctions you know what will happen  . My home get even fired because of it. It will be very useful feature for most of us Thank you very much in advance! Best |
|
|
|
And you'll order 1 additional board per 6 planed now . This is a good thing for us assemblers . I'm joking , of course true marto unfortunately you are correct |
|
|
|
88% of load is too much, personally I will not load voltage regulator with more than 70%, ye we all think of the lower cost, lower space but adding some regulators for 10$ more will help a lot in safety and reliability, this change will increase price of the product with arround 5% which is funny.
+1 But let us leave regs for now. When BKK reach the "first share point" and have the board hashing he can easily measure the chip consumption itself at 350 and share the result. Above 350 is not needed as long AS avalons are not performing and producing a lot of crap HW errors. then we can move on with that discussion. If board stay unchanged in terms of power design i will just remove 2 or 4 chips and that is all. if i do have 12 chips the planned regs will be driven at 75% max. |
|
|
|
---------------
2. Scrub the extra power connector behind the PCIe and add a IR3897 9A reg there to supply 4 of the end chips. This would cost a bit extra for parts - rough guess, about $5 maybe.
--------------
+1, 4 chips x 2A = 8A and we will drive IR3897 at 88% load. Can we just solve one problem and create another? Do we need powerful reg? 10 Amps at least? |
|
|
|
Adding additional voltage regulators will help not only for the overclock but for more stable work, lower noise psu, lower temperatures, if you make board 12X12 this will help also about cooling, cause the most produced fans are with this size, now stocking a quallity 10cm fan is hard.
You are damn right about it. But it costs money for sure. The goal is to make it work in a stable way and to minimize the cost. this is just a compromise that we have to make and reach our sweet spot |
|
|
|
Unfortunately when I started the design I hadn't heard anything about over clocking Avalons and didn't design that in. I added some head room on the power supply just out of safety. I'm not aware of any higher current easy to swap regulator solution so the 3 options I see are:
1. Just place 14 chips / board. The wasted 2 chip board space isn't much of a loss.
2. Scrub the extra power connector behind the PCIe and add a IR3897 9A reg there to supply 4 of the end chips. This would cost a bit extra for parts - rough guess, about $5 maybe.
3. As #2 but extend the board by about 2cm and add 4 more chips, plus a full 16A IR3895 reg behind the PCIe conn. This would supply 6 chips, with 2 of those taken from current 16, making a 20 chip board. This would cost only a bit more than #2 due to higher board cost and the parts being a bit more costly, rough guess, about $8 maybe, but a more radical design change due to heat sink size also being different.
Another data point - the IR3895 over current protection kicks in at 18A minimum, so how well it behaves at max load probably depends a lot on how well it's cooled. Also, some of the reported power use in the other thread is for control board, 3.3V supply and fans so it could be somewhat less for actual ASIC power. Same performance with 14 clocked chips like 16 not clocked. And you save two chips and the end
What would be the (estimated) hash rate of those 3 options? I think Option #3 is probably too radical especially if folks like Terrahash have cases being built Depends how much are you gonna to overclock them. In general with 14 chips you are loosing 2 chips = 13% of the hash power (300 clock), but you gain 15% with overclocking to 350 per chip which is at + for me |
|
|
|
the stromborn version has this values
case "$_cf" in
256)
CF="50:256"
;;
270)
CF="47:270"
;;
282)
CF="45:282"
;;
300)
CF="40:300"
;;
305)
CF="39:305"
;;
310)
CF="38:310"
;;
315)
CF="38:315"
;;
320)
CF="37:320"
;;
325)
CF="36:325"
;;
330)
CF="36:330"
;;
335)
CF="35:335"
;;
340)
CF="35:340"
;;
345)
CF="34:345"
;;
350)
CF="34:350"
;;
355)
CF="33:355"
;;
360)
CF="33:360"
;;
365)
CF="32:365"
;;
370)
CF="32:370"
;;
375)
CF="32:375"
;;
380)
CF="31:380"
;;
385)
CF="31:385"
;;
390)
CF="30:390"
;;
395)
CF="30:395"
;;
esac
my sweet spot is 355 but I have only 2 modules Avalon V1
I think that the hw rate is dependent on the controller + temp
I had 0,4% increase in hw with 370 but I have already a broken module so 355 is better
They will not work with the default official firmware if you just insert them inside. Tried and was unable to get some configs, I am assuming the firmware has to support certain speeds like 345, 355, etc Btw, how did you break a module ? |
|
|
|
the stromborn version has this values
case "$_cf" in
256)
CF="50:256"
;;
270)
CF="47:270"
;;
282)
CF="45:282"
;;
300)
CF="40:300"
;;
305)
CF="39:305"
;;
310)
CF="38:310"
;;
315)
CF="38:315"
;;
320)
CF="37:320"
;;
325)
CF="36:325"
;;
330)
CF="36:330"
;;
335)
CF="35:335"
;;
340)
CF="35:340"
;;
345)
CF="34:345"
;;
350)
CF="34:350"
;;
355)
CF="33:355"
;;
360)
CF="33:360"
;;
365)
CF="32:365"
;;
370)
CF="32:370"
;;
375)
CF="32:375"
;;
380)
CF="31:380"
;;
385)
CF="31:385"
;;
390)
CF="30:390"
;;
395)
CF="30:395"
;;
esac
my sweet spot is 355 but I have only 2 modules Avalon V1
I think that the hw rate is dependent on the controller + temp
I had 0,4% increase in hw with 370 but I have already a broken module so 355 is better
They will not work with the default official firmware if you just insert them inside. Tried and was unable, I am assuming the firmware has to support certain speeds like 345, etc Of course they will not work. you need to run patched cgminer. strombom could you please post the patch? It will be of great use for some of us |
|
|
|
Unfortunately when I started the design I hadn't heard anything about over clocking Avalons and didn't design that in. I added some head room on the power supply just out of safety. I'm not aware of any higher current easy to swap regulator solution so the 3 options I see are:
1. Just place 14 chips / board. The wasted 2 chip board space isn't much of a loss.
2. Scrub the extra power connector behind the PCIe and add a IR3897 9A reg there to supply 4 of the end chips. This would cost a bit extra for parts - rough guess, about $5 maybe.
3. As #2 but extend the board by about 2cm and add 4 more chips, plus a full 16A IR3895 reg behind the PCIe conn. This would supply 6 chips, with 2 of those taken from current 16, making a 20 chip board. This would cost only a bit more than #2 due to higher board cost and the parts being a bit more costly, rough guess, about $8 maybe, but a more radical design change due to heat sink size also being different.
Another data point - the IR3895 over current protection kicks in at 18A minimum, so how well it behaves at max load probably depends a lot on how well it's cooled. Also, some of the reported power use in the other thread is for control board, 3.3V supply and fans so it could be somewhat less for actual ASIC power.
As i said K14 is the easiest path. thank you for the detailed update BKK! |
|
|
|
Do you think that power regs and other parts in K16 will be able to stand the high power consumption due to overclocking?
Going by the numbers it looks borderline maximum. How well it works in reality, I don't know yet. He reports 350MHz draws 576W PSU output. With 240 chips thats 2.4W/chip or 2A @ 1.2V. For 8 chips/bank that's the rated limit of the regulator, 16A. That are bad news. Is there more powerful reg that can be used? If not I will go with K14 |
|
|
|
Bump.
Still nothing now.
A lot of us still did not this applies to trade-ins also. Patience dude |
|
|
|
Bkk, With latest overclocking info and measurements, https://bitcointalk.org/index.php?topic=140539.msg2552912#msg2552912Measured at the wall (plug) it is about 10% increase from 300 to 350. Do you think that power regs and other parts in K16 will be able to stand the high power consumption due to overclocking? I will be very glad when you reach the hashing point to measure Avalon chip consumption at 350 My personal oppinion is that you should consider the overclocking in this stage of development and change some elements on the fly if needed for K16 Pls comment 10X |
|
|
|
|
ebereon,
Great post about power!
What about hash rate at 375?
I am afraid to test it and as reported there is no big increase performance
|
|
|
|
Blue fan pulls air into unit if I understand correctly
Yes you do i have just checked it and it is indeed. i get confused 10X |
|
|
|
So how hot is too hot?
I'm running my batch 1 and batch 2 units at 350MHz now. B1 seems to have settled on 47c for temp3, where B2 is at 52c now. Ambient temperature here is somewhat high (28c).
I have to place my unit right at the mouth of AC vent to make sure temp3 < 50C. In SW USA, top floor Dude, Can you tell me batch one unit blue fan is it sucking or blowing cold air like PWMS? Please take a look at my last post |
|
|
|
|
Show Posts
