A fool and his not-really-money are soon parted
-
@Luhmann said in A fool and his not-really-money are soon parted:
@Lorne-Kates said in A fool and his not-really-money are soon parted:
Excel spreadsheet is to a game of Tetris.
I guessed it had been done ... but appears to have been done a dozen times.
But can
you make a pivot table of high scoresit run Doom?
-
@JBert Oh god. I don't think I'd be surprised if someone did render 3D graphics in Excel. It's just matrix transformations at the end of the day, right?
-
-
@Luhmann said in A fool and his not-really-money are soon parted:
@Lorne-Kates said in A fool and his not-really-money are soon parted:
Excel spreadsheet is to a game of Tetris.
I guessed it had been done ... but appears to have been done a dozen times.
The first time I saw something like this would have been in the year 2000, or maybe 2001, and the game was Pacman, so Tetris doesn't surprise me at all.
-
@pie_flavor Interesting about mirroring the top and bottom.
I sure hope nobody unrolls loops so drastically these days.
As an aside, it's odd being able to quantify the processor clock speed into a finite, certain amount of instructions. You don't get those sort of liberties beyond assembly.
-
@kazitor said in A fool and his not-really-money are soon parted:
@pie_flavor Interesting about mirroring the top and bottom.
I sure hope nobody unrolls loops so drastically these days.
As an aside, it's odd being able to quantify the processor clock speed into a finite, certain amount of instructions. You don't get those sort of liberties beyond assembly.
You don't get that liberty in assembly these days, because of things like:
- SpeedStep and similar changing the clock frequency and its relationship to the memory clock frequency.
- Thermal braking slowing the processor and changing its relationship to the memory clock frequency.
- Variations in the cycles taken to fetch data because of caching.
- Variations in the cycles taken to fetch instructions because of caching.
- Variations in the above because of TLB misses.
- Variations in the above because of unaligned data and/or instructions.
Um. Just to be clear: those things mean that you cannot say "in X clock cycles I can execute Y instructions". The cache-related things also mean that you can't say exactly how many cycles it will take to fetch and completely execute an instruction, never mind the fun and games of out-of-order and parallel execution.
-
@Steve_The_Cynic said in A fool and his not-really-money are soon parted:
Just to be clear: those things mean that you cannot say "in X clock cycles I can execute Y instructions".
That depends on the architecture, and how completely the application has control. Embedded systems can say that sort of thing…
-
-
@Luhmann Oh, it's the old version of Ethereum, which kept going after a larger group of people voted to undo an earlier attack and thus abandon the previous blockchain state.
-
@JBert said in A fool and his not-really-money are soon parted:
@Luhmann Oh, it's the old version of Ethereum, which kept going after a larger group of people voted to undo an earlier attack and thus abandon the previous blockchain state.
@TheFuckingArticle says:
Stated a different way, a rollback attack generates a new fork of the blockchain.
So a blockchain fork was forked-by-attack and now they'll have to fork the fucked fork to get fully functional again.
@ThatFreakingArticle says:
ETC officials, for their part, have confirmed that double spends are affecting the currency, but they have yet to say more.
"affecting" meaning the value dropped from ~$5.20 to ~$4.80. Yeah, I remember that happens with traditional currency when there's a bank robbery and the USD suddenly loses 8% of it's value.
Stable!
-
@dkf said in A fool and his not-really-money are soon parted:
@Steve_The_Cynic said in A fool and his not-really-money are soon parted:
Just to be clear: those things mean that you cannot say "in X clock cycles I can execute Y instructions".
That depends on the architecture, and how completely the application has control. Embedded systems can say that sort of thing…
You're not totally wrong, except that merely saying "embedded systems" isn't sufficient. There have been embedded systems with processors with e.g. caching for a long while now (the first one like that I worked with was a MIPS R4600 in 1995(1)). Anything with x86 in it has suffered from variability due to the unaligned instructions thing ever since the 8086. (Curiously, the 8088/188/288 didn't suffer from that, because all instructions were aligned on 8-bit boundaries, obviously.)
(1) MIPS processors, it turns out, cannot automatically load their own TLB entries, so a TLB miss (not a page fault, but a TLB miss) raises an exception. We ran with paging on, but with a fixed page table that mapped the discontiguous physical address ranges of the RAM into a contiguous logical space.
-
@Steve_The_Cynic said in A fool and his not-really-money are soon parted:
Anything with x86 in it has suffered from variability due to the unaligned instructions thing ever since the 8086.
If you're serious about this stuff, you know the alignment of the instructions.
-
@dkf said in A fool and his not-really-money are soon parted:
@Steve_The_Cynic said in A fool and his not-really-money are soon parted:
Anything with x86 in it has suffered from variability due to the unaligned instructions thing ever since the 8086.
If you're serious about this stuff, you know the alignment of the instructions.
For sure, but it's more important to note that in order to analyse the count of clock cycles, you need to know the alignment of the instructions. Well, and whether they cross cache line boundaries, page boundaries, megapage boundaries(1), etc. and whether at a particular moment the cache line(s) and page table entry(-ies) are in the CPU or whether it has to go out and fetch them from physical memory. (That is, the alignment of the instructions is almost irrelevant on modern (say, after the Pentium Pro in 1995) CPUs.)
There are applications where locality of reference just plain does not happen in the data set, even if the code's own locality is significant. Based on something a colleague said once about some work a friend of his was doing, finite element analysis suffers greatly from this, and the widely distributed nature of the data set meant that almost every reference to the finite elements would cause a total cache miss, and the CPU's habit of fetching an entire cache line on a cache miss meant that the job ran at about half the speed on a P4 (128-byte cache line) that it did on a P3 (64-byte cache line) of the same FSB speed. In such applications, only the memory speed parameters matter.
Summary: Estimating the speed of / time required for a computing job is hard, and modern systems have done a great deal to make it harder.
(1) That is, the instruction is located in two pages, and those pages' page table entries aren't in the same page table, so the CPU must go back to the page table directory to find the address of the other page table. It's ... different on PAE-enabled x86 processors with PAE active, which of course includes everything in 64-bit mode.
-
@Steve_The_Cynic said in A fool and his not-really-money are soon parted:
For sure, but it's more important to note that in order to analyse the count of clock cycles, you need to know the alignment of the instructions. Well, and whether they cross cache line boundaries, page boundaries, megapage boundaries(1), etc. and whether at a particular moment the cache line(s) and page table entry(-ies) are in the CPU or whether it has to go out and fetch them from physical memory. (That is, the alignment of the instructions is almost irrelevant on modern (say, after the Pentium Pro in 1995) CPUs.)
You don't tend to need to know that on embedded CPUs; they don't really use paging all that much. Moreover, when you're using an ARM processor derivative, the instruction alignment is a heck of a lot simpler anyway.
-
@dkf I think the most complicated it gets on ARM are certain sub-architectures which use both 16-bit and 32-bit instructions intermixed?
-
@PleegWat said in A fool and his not-really-money are soon parted:
I think the most complicated it gets on ARM are certain sub-architectures which use both 16-bit and 32-bit instructions intermixed?
Yes… but you don't normally do that inside a function as it involves twiddling with the CPU status register to switch it between the modes.
-
@dkf Ah, I didn't know that. So some potential for mis-interpreting instruction data as 32-bit instead of 16-bit or vice versa.
-
@PleegWat said in A fool and his not-really-money are soon parted:
some potential for mis-interpreting instruction data
Yes, but easier to just jump into the middle of your actual data if you want trouble, assuming you're not using a processor with a memory protection control module (they're common in systems with modern OSs, but rare in the embedded space where very different hardware modules are used instead).
-
@kazitor said in A fool and his not-really-money are soon parted:
@JBert Oh god. I don't think I'd be surprised if someone did render 3D graphics in Excel. It's just matrix transformations at the end of the day, right?
@JBert said in A fool and his not-really-money are soon parted:
But can
you make a pivot table of high scoresit run Doom?
-
-
@DCoder Not sure if stupid because crypto or brilliant because “Fuck
youTrump, give me money!”
-
@DCoder That's every headline at once!
-
@DCoder said in A fool and his not-really-money are soon parted:
LOL:
The estate describes a major drawcard of ESCOBAR to be that most stablecoins have been launched by “unknown companies,” as opposed to Escobar Inc.
-
@boomzilla There's a good joke about that, but I don't remember it fully and can't find the right quote. Might be something Douglas Adams-y, maybe someone can remind me again. It goes along the lines of:
It's not the liars who are problematic, you can always count on them to lie to you, but the honest people where you can never be sure.See post below.In this sense, that's a very calculated risk, because you can pretty much count on "Escobar Inc." trying to rip you off.
-
@topspin
It's probably not uncommon in many works said in different ways, but what I recall best is how Jack Sparrow sells it in POTC:"Me? I'm dishonest, and a dishonest man you can always trust to be dishonest. Honestly. It's the honest ones you want to watch out for, because you can never predict when they're going to do something incredibly... stupid."
-
@Applied-Mediocrity The source is very far from what I guessed but yes, that’s what I had in mind. Thanks.
-
Well that does it, iFramely must be disabled.
-
-
@Tsaukpaetra Oh, now I get it
-
@hungrier Can you please enlighten me?
-
@kazitor Those are screwdriver bits inserted into holes in coins.
-
@loopback0 I still don't get it. Is there a pun involved or not really?
-
@Zecc Coins with bits in. Bit coins.
-
@loopback0
Fsck! Fsck! My tiny brain! Goddammit... AAAARGH! >_<
-
Oh hai guyz, GPU mining is still a profitable business, right? Right?!
While most mid/high-end GPUs have at least 8GB of memory, very few have more than 12GB, and 16GB will be required to mine using the 2020 variant of Grin Mean AF Cuckatoo algorithm. A GPU with a 16GB frame buffer could conceivably mine both the GPU friendly and ASIC friendly algorithms for more than 2 years whereas others are limited to just the GPU friendly Cockaroo after the first year. The money printing machine.
-
@Atazhaia said in A fool and his not-really-money are soon parted:
The money printing machine.
-
@Luhmann said in A fool and his not-really-money are soon parted:
@Atazhaia said in A fool and his not-really-money are soon parted:
The money printing machine.
"Shitcoins are the future of currencies."
"Mining shitcoins is a money printing machine."
I don't see any problems with that at all.
-
@Atazhaia said in A fool and his not-really-money are soon parted:
The money printing machine.
It's a video card and printer combo?
-
@TimeBandit
It's probably an InktJet or similar so the cartridges cost are YUGE and they dry out if you stop printing for a day
-
@Atazhaia said in A fool and his not-really-money are soon parted:
Grin Mean AF Cuckatoo algorithm
Is that really what it's called?
-
@loopback0 said in A fool and his not-really-money are soon parted:
@Zecc Coins with bits in. Bit coins.
Geez. I just couldn't see that. And now I feel really stoopid. As I should.
-
@hungrier said in A fool and his not-really-money are soon parted:
@Atazhaia said in A fool and his not-really-money are soon parted:
Grin Mean AF Cuckatoo algorithm
Is that really what it's called?
Considering that it's cryptocurrencies we're talking about, it wouldn't surprise me...
-
@dcon said in A fool and his not-really-money are soon parted:
@loopback0 said in A fool and his not-really-money are soon parted:
@Zecc Coins with bits in. Bit coins.
Geez. I just couldn't see that. And now I feel really stoopid. As I should.
Now tell us: what do you think of Bitcoin? No wrong answers, we're establishing baseline, but do try to answer honestly as we'll be using these results in figure tests.
-
@Tsaukpaetra said in A fool and his not-really-money are soon parted:
@dcon said in A fool and his not-really-money are soon parted:
@loopback0 said in A fool and his not-really-money are soon parted:
@Zecc Coins with bits in. Bit coins.
Geez. I just couldn't see that. And now I feel really stoopid. As I should.
Now tell us: what do you think of Bitcoin? No wrong answers, we're establishing baseline, but do try to answer honestly as we'll be using these results in figure tests.
You mean this?
-
@dcon Did you ask before posting a personal photo of him?
-
-
@dcon I can see what looks like a speaker grille; does it say "Parp!" when you put in coins?
-
@Atazhaia said in A fool and his not-really-money are soon parted:
The money printing machine.
It's true. Selling overpriced hardware to gullible coin-hunting idiots IS a money printing machine.
-
@Lorne-Kates said in A fool and his not-really-money are soon parted:
@Atazhaia said in A fool and his not-really-money are soon parted:
The money printing machine.
It's true. Selling overpriced hardware to gullible coin-hunting idiots IS a money printing machine.
During a gold rush, the best money is to be made selling pickaxes.
-
@PleegWat said in A fool and his not-really-money are soon parted:
During a gold rush, the best money is to be made selling pickaxes.