Progress on Blunder

[j.t.]

...

I don't like to use all of the SMT/HyperThreading threads for chess engines.

I haven't thought about that. Does hyperthreading not work well for chess engine matches?

[mvanthoor]

I haven't thought about that. Does hyperthreading not work well for chess engine matches?

I don't know. Some people are adamant that it doesn't work and shouldn't be used for matches; they even go as far as to turn hyperthreading completely off in the BIOS to be sure no engine ever gets assigned to such a thread. I can understand why. Some even turn off all forms of speed-step and turbo-boost, but that's going a step too far for me. I'm not going to run my CPU at full speed all of the time.

Other people actually use these threads as cores; so if they have an 8-core machine with hyperthreading, they run 16 games.

I'm somewhere in the middle. The 6700K is my main system right now, so if I run a test, I run 4 games concurrently. I use the 4 HT-cores for things such as browsing and watching a YouTube video, that sort of thing. After I build the 5700G, which will be a dedicated chess / test computer for as long as I actively develop my engine, I'm probably going to run 12 games concurrently, leaving 4 of the HT/SMT threads for looking in on those games through VNC and such.

I just feel it's not a good idea to completely slam a system with the maximum number of engines because it won't have any power left to do anything else. As soon as something happens somewhere (a task gets run, you log in, etc...) some engine is going to get disadvantaged, and the task that is run will probably be slow as molasses. Therefore I'll always leave some of the HT threads unused.

I have been thinking to just build a Zen3 system around the 5950X and run 24 or 28 matches concurrently, but I think that's a bit over the top. My engine is not yet far enough along to warrant that, and in 2023 I'll probably going to build a main desktop anyway. That'll also be at least a 16-core machine that'll then be using for around 8 years. (6-7 years is the point where I start thinking about a new computer, and the 8 year mark is where I typically upgrade, at least nowadays.)

[algerbrex]

Just thought I'd give a quick update, seeing as it's been awhile.

As I've mentioned in other places, I'm currently rewriting/refactoring Blunder from scratch, to improve speed, design, and to use a much more rigorous testing framework to ensure consistent and quality progress in this newer version. No more being lazy and accepting a patch because it's "good enough." If it's not proven to gain Elo, or doesn't contribute to the design or personality of Blunder in a way I deem useful, it's getting chucked.

With that said, all progress on Blunder has currently been halted, as my life has gotten much busier this second semester of Freshmen year of college. Between keeping up with classwork, working part-time, maintaining relationships, and doing extracurriculars, I simply don't have the time to sit down and work on Blunder for long hours as I use to.

I haven't lost interest in developing Blunder further, or computer-chess programming, or chess in general for that matter. Although I'd be lying if I didn't admit it has waned a bit. But as of right now, I'm not sure when I'll pick Blunder up again. Perhaps not until another couple of months when the summer comes around, although I anticipate it'll be much earlier than that.

If in the coming months I do realize I'm no longer interested in working on Blunder, I'll make another post detailing that separately.

[Chessqueen]

Just thought I'd give a quick update, seeing as it's been awhile.

As I've mentioned in other places, I'm currently rewriting/refactoring Blunder from scratch, to improve speed, design, and to use a much more rigorous testing framework to ensure consistent and quality progress in this newer version. No more being lazy and accepting a patch because it's "good enough." If it's not proven to gain Elo, or doesn't contribute to the design or personality of Blunder in a way I deem useful, it's getting chucked.

With that said, all progress on Blunder has currently been halted, as my life has gotten much busier this second semester of Freshmen year of college. Between keeping up with classwork, working part-time, maintaining relationships, and doing extracurriculars, I simply don't have the time to sit down and work on Blunder for long hours as I use to.

I haven't lost interest in developing Blunder further, or computer-chess programming, or chess in general for that matter. Although I'd be lying if I didn't admit it has waned a bit. But as of right now, I'm not sure when I'll pick Blunder up again. Perhaps not until another couple of months when the summer comes around, although I anticipate it'll be much earlier than that.

If in the coming months I do realize I'm no longer interested in working on Blunder, I'll make another post detailing that separately.

It will take a few years , from the time that Blunder makes no more Blunders and you will have to change the name to Perfect

[Chessqueen]

Just thought I'd give a quick update, seeing as it's been awhile.

As I've mentioned in other places, I'm currently rewriting/refactoring Blunder from scratch, to improve speed, design, and to use a much more rigorous testing framework to ensure consistent and quality progress in this newer version. No more being lazy and accepting a patch because it's "good enough." If it's not proven to gain Elo, or doesn't contribute to the design or personality of Blunder in a way I deem useful, it's getting chucked.

With that said, all progress on Blunder has currently been halted, as my life has gotten much busier this second semester of Freshmen year of college. Between keeping up with classwork, working part-time, maintaining relationships, and doing extracurriculars, I simply don't have the time to sit down and work on Blunder for long hours as I use to.

I haven't lost interest in developing Blunder further, or computer-chess programming, or chess in general for that matter. Although I'd be lying if I didn't admit it has waned a bit. But as of right now, I'm not sure when I'll pick Blunder up again. Perhaps not until another couple of months when the summer comes around, although I anticipate it'll be much earlier than that.

If in the coming months I do realize I'm no longer interested in working on Blunder, I'll make another post detailing that separately.

It will take a few years , from the time that Blunder makes no more Blunders and you will have to change the name to Perfect

Or several years until you become very Old, from the time that Blunder makes no more Blunders and you will have to change the name to accuracy

[algerbrex]

I was going to wait to share this logo when I released Blunder 8.0.0 (hopefully in the next week or two), but I decided I'd go ahead and share it now. It's pretty simple, something I whipped up in Gimp in about 15-20 minutes, but I like it and think it suits Blunder pretty well!



I had the idea for the logo a while back last, but only recently go around to making it.

[mvanthoor]

I had the idea for the logo a while back last, but only recently go around to making it.

Good idea, but the kerning and boldness are off. Therefore it now reads "B+ Lunder".

Personally I'd try something like this (also removing the shadow):



(Obviously this isn't perfect either, because the B is soft now and the rest of the letters aren't even bold enough yet, but I GIMP-ed this in about 5 minutes. I'm sure you get the idea.)

Even my own logo (here) has a tiny kerning problem between the R and U, and the i and c. "ustic" should be moved closer to the R by maybe 10 or 15 pixels; same for the c towards the i. It bugs me to no end. At some point I have to fix this. (If I remember correctly I still have the original XCF file from GIMP on my Windows-partition.)

Since I started working with LaTeX 10 years ago and learned how kerning and typography should be done properly, I'm seeing kerning and other typography weirdness everywhere and I can't unsee it

[algerbrex]

Good idea, but the kerning and boldness are off. Therefore it now reads "B+ Lunder".

Personally I'd try something like this (also removing the shadow):



(Obviously this isn't perfect either, because the B is soft now and the rest of the letters aren't even bold enough yet, but I GIMP-ed this in about 5 minutes. I'm sure you get the idea.)

Ah, so that's what the spacing between letters is called. Learn something new every day.

Anyhow, thanks for the feedback. I can see what you mean. I suppose I'm a little biased since I was looking at the logo and design for so long to figure out the details, I can read the name "Blunder" fairly quickly. But objectively speaking the spacing is off.

I recognized the spacing issue a little bit when I first made the logo, because of the cross at the top of crown, but I wasn't entirely sure how I wanted to address it. The more I look at your tweaks though, the more I like the design. I'll open back up the originally XCF file and see if I can't make it look a bit more aesthetically pleasing.

As far as the boldness goes, I remember I wanted to make the first letter "pop" of course, but I think I did go a little overboard and the shadow was a bit of an afterthought, because in my mind the logo by itself looked a tad bland, and I was looking for another way to make it stand out more. I'll probably revisit some ideas for this too.

Even my own logo (here) has a tiny kerning problem between the R and U, and the i and c. "ustic" should be moved closer to the R by maybe 10 or 15 pixels; same for the c towards the i. It bugs me to no end. At some point I have to fix this. (If I remember correctly I still have the original XCF file from GIMP on my Windows-partition.)

Since I started working with LaTeX 10 years ago and learned how kerning and typography should be done properly, I'm seeing kerning and other typography weirdness everywhere and I can't unsee it

I'll be honest, I've never noticed a spacing issue with your logo, and I still can't really see it. The curse of typographist I suppose This is why I eventually plan to make my living from being a software engineer and not a graphic designer.

[mvanthoor]

I recognized the spacing issue a little bit when I first made the logo, because of the cross at the top of crown, but I wasn't entirely sure how I wanted to address it. The more I look at your tweaks though, the more I like the design.

Thanks. The reason why I slanted the "lunder" part of the logo is because if you keep it upright, the spacing between the bottom of the "B" and the "l" is still to large, even if you move the L over as close as it can go without colliding with the bend of the lower part of the B. By slanting the L, the lower part of that letter can move closer to the B, while keeping the middle part about the same distance from it.

I'll be honest, I've never noticed a spacing issue with your logo, and I still can't really see it. The curse of typographist I suppose This is why I eventually plan to make my living from being a software engineer and not a graphic designer.

As I said, the kerning problem on my logo is tiny, and indicative of the font I used. The "u" is a bit too far to the right; because the bend curves inward, the spacing between the bottom right foot of the "R" and the curve on the left of the "u" is a bit too large. Same problem on the bottom foot of the "i" and the curve of the "c". The U and C maybe need to move over by only 10 pixels or so to fix it.

When I did that LaTeX tutorial, it pointed out some notable differences between LaTeX kerning / typography and the lack thereof in Word (at the time). Because of that I looked into typography some more. I know just enough to be dangerous and see some common or obvious imperfections here and there. Don't ask me to design a font though.

[algerbrex]

As I said, the kerning problem on my logo is tiny, and indicative of the font I used. The "u" is a bit too far to the right; because the bend curves inward, the spacing between the bottom right foot of the "R" and the curve on the left of the "u" is a bit too large. Same problem on the bottom foot of the "i" and the curve of the "c". The U and C maybe need to move over by only 10 pixels or so to fix it.

Ah, I see a bit more of what you mean.

When I did that LaTeX tutorial, it pointed out some notable differences between LaTeX kerning / typography and the lack thereof in Word (at the time). Because of that I looked into typography some more. I know just enough to be dangerous and see some common or obvious imperfections here and there. Don't ask me to design a font though.

Well if I do end up going down an applied math right like I am right now, I'm sure I'll be encountering a lot more LaTex, and I'll get an eye for these sorts of these things

By the way, I noticed you hadn't had an opportunity to release Rustic 4 yet.