Thanks. Yes, all 7 are 3d stable into the upper 1300s megahertz. Four of them have 1332mhz boost clocks, two have 1345mhz, and one is 1319mhz. Voltage varied from 1.137 to 1.174v. The stock +35% power target increase is enough to not perf. cap them on power.

Three seemed to be 3d stable at 1.4ghz with the best going up to 1.44ghz at only 1.137v.

I didn't individually check the memory overclocks which are actually more important for performance. At least one in my system doesn't like +600 so I've been using +550 for now.

The systems cost comes out to $1,542 or 1.22 Kh/USD. 4gb and a smaller psu would have saved a little but I want my smaller gpu rigs to be ready to add another 1000w psu and be able to run six beefier cards in the future. If its still economical to do so.

Here are my initial results with six 750 Ti FTWs at +40 core (1360-80mhz-ish) and +550 memory.



81w idle and 547w mining scrypt. 3.46 K/hash per watt.

Powered usb 3.0 risers, MSI Z77A-GD65, 1000w Raidmax gold psu.

I tried to use 7 cards but my system won't even post with it physically plugged into in the last slot.

Overall, a nice inexpensive upgrade from my three 670s. The default 42% fan speed noise level is ridiculous though.

I have a 750 Ti FTW at 1363mhz core (+40) and 1600mhz memory (+500) with these results:

330 K/hs scrypt (-H 1 -l T5x24)
37w idle
120w load

The other components are: 450w gold, G1610, Asrock extreme4, 2x 4gb 1.5v, 80gb ssd, gigabit link, usb bt m/kb, on 120v.

I can also confirm the bios has a built in 35% power limit increase like advertised. It shows a 52w "TDP" value instead of the default 38.5w. With this +35%, the card is more voltage cap limited than power limited. I haven't been able to get 1.2v modded like StayPuftOCN did with his Asus card. My 750 Ti FTW uses 1.15 volts and it isn't 1.4ghz 3d stable. 

Do it in dos with the provided nvflash.exe. Worked for my 750 ti ftw, which doesn't really need more power limit though since it's vOP (operating voltage) limited in all cases as is.

The windows and os x wallets were posted before the launch. I got some blocks in the first minute while running the windows wallet so I don't understand why I keep seeing this claimed?

There was only like 50 blocks in the first minute. Also, I think we were a little over block 200 after the first hour so the initial rush added less than an hours worth of coins.

Are you running the wallet on the pc you're mining on? That really hurt my utilization when I tried that momentarily.

I'm not sure if it was posted in here but here's a link to a post with some commands to check your wallet's balance and other network information.

https://bitcointalk.org/index.php?topic=434190.msg4863787#msg4863787

Yes, unless it ends up an orphan.

I should go buy a lottery ticket.... thanks Christian!

Earlier today someone posted this update:

Errm, 1/5 the maintenance of Litecoin with less than 1/50th the market cap?

Yaccoin is still faster using cudaminer and its already using the n-factor (32768) that Vercoin will be on in 2022.

I had to go from 4gb to 8gb of RAM to get my 4th gpu to work. But you probably have 4gb already so I don't know...

There's one like that on my pool too. Had over 100k KH/s earlier for an estimated 1300 coins a day.

From the readme:

"--single-memory  [-m] list of flags (0 or 1) to make the devices
                 allocate their scrypt scratchpad in a single,
                 consecutive memory block. On Windows Vista, 7/8
                 this may lead to a smaller memory size being used.
                 When using the texture cache this option is implied."

"--hash-parallel  [-H] scrypt also has a small SHA256 component to it:
                      0 hashes this single threaded on the CPU.
                      1 to enable multithreaded hashing on the CPU.
                      2 offloads everything to the GPU (default)"

I have the most efficient results with m1 and h2 with my 780 ti in windows.

Using the Z kernel I just broke 900 khash/s with a water cooled 780 TI. This kernel ended up being 11% faster than the 12-18 releases T kernel. I had to use -H 1 for the extra 8 hashes over -H 2, at the cost of 25w. A terrible trade, but it put me over 900.

N-factor 9 Microcoin is running at @ 834 khash/s. That's a much smaller loss than a R9 290 takes. They are now yielding 666 vs 880 on scrypt. The memory overclock actually helps on Microcoin, but it isn't an efficient trade. 8~ hashes for 8~ watts. Also, solo mining MRC was terrible for me in the few hours I tried it. I had several times better results in the blocksolved pool.