DIARY 2013














Sun, Aug 18

<18:40 EEST> The summer is almost over for some people, but mine is only halfway through in some sense. Olviretki Schleusingenissa and Makeen laakson lukio have plenty of shows left, but right now I am enjoying a few random days off the stage.

The planned hiatus was highlighted by a preceding week+ with a show every night and a summer school course during the days. For some reason I found myself doing teacher training of sorts, having a general interest in the matters and having a free pass as a current student. Now it is time to play with some hardware and software, and I have kept myself busy with the aforementioned random number generator — something I guess every electro-hacker should do once in a lifetime. In fact, I was looking for a straightforward solution to improve the /dev/random for a specific software application, but having failed the search, I decided to roll my own, and it has been loads of fun. I could even recycle some of my Bitcoin mining cluster ideas for a multi-port setup, albeit with an inverted sense of clustering many computers with one FPGA.

The software project in question is Primecoin. For once, I managed to follow an interesting altcoin in advance, and start mining from day 1. Indeed, this coin (aka XPM, presumably to cater for the XAU/XAG/XRP standards) has interesting mining aspects because it is still CPU-only. The underlying reasons are all the more intriguing: the proof of work deals with prime numbers, their search being of the few reasons I have ever bought excessively powerful hardware.

Alas, anyone into prime search should know that new prime discoveries are huge, and cannot possibly be verified within the minute that is the block interval. Instead, Primecoin searches for Cunningham chains of primes. Nevertheless, it is an ingenious way to combine a scientifically interesting work into cryptocurrency transaction verification. That said, some people may still find this wasteful — they should look into Peercoin (PPC), which is also one of my main interests in the scene.

Speaking of CPUs and primes, I have also strengthened my position with regards to GIMPS, including trial factoring on AMD GPUs. I have also done more benchmarking on CPUs and made a rather odd discovery: my old Core 2 Duo T7200 is much faster than the new laptop's Core i5 M520, when it comes to the LL test for new Mersenne primes. And not just a little faster, but 70% so, despite having a slower clock at 2.00 vs. 2.40 GHz. (There is probably an issue of cache usage -- T7200 has 4 MB of shared L2, while M520 has 2x 256 KB of L2 and 3 MB of shared L3.) Arguably, the new CPU is probably faster in some other applications, particularly if they can use SSE4. Too bad the new CPU does not have AVX, it would double the LL performance :-/

Well, perhaps a few words about the healthy outdoor life in the summer theatre are in order at this point. The JYT piece is a typical indoor production, though, but it has not consumed my time and soul like the other play — in retrospect, Makeen laakson lukio turned out surprisingly easy for me, totally routine, though the live mic effects (using my Zoom H4n) do add a little excitement.

Olviretki has been a different beast altogether. I started out presenting myself as a music/sound guy, and agreed to do a small supporting role in addition. As time went by, we could not find enough people for a few other roles, so I now play one rather major character, in addition to two others, resulting in rather tight squeezes of changing clothes on the fly. Oh, and I still manage most of the music and sound. The lack of staff means that a few actors have to play the sound guy too — again powered by my trusted Zoom. Things get seriously fun when it's raining and I have to set up the sound system, besides preparing my sets of clothes and all.

I do mean fun, though. About 7 years after my last proper role, acting is great fun, even if my main artistic talent is recalling long lines to the letter. I just wish I could really focus on it. Playing too many stage and technical roles at once feels like doing a half-assed job in all of them. Not that I am alone, the lack of staff is a general issue. It is probably the worst for the director who cannot see his first major work with his own eyes — as he is playing my mate.

Risto A. Paju