josmala

4080 12GB did its task. It send the gamers a message that do not wait for cheaper 40 series cards go by 30 series now. Now the best thing Nvidia can do is wait until navi 3 has launched and release 4070 ti with same specs but decide price point based on what AMD has set to their cards.

That 12GB 4080 had compute bit above 3090 ti. But memory bandwidth between 3070 and 3070 ti.

If it the cache hit rate is similar to Navi2 of similar size. It would have effective bandwidth similar to 3090 in 1440p and way above any previous gen card in 1080p but at 4k it would be around 3070 ti territory in terms of memory bandwidth. So a great 1440p card natively and to those who are willing to use DLSS to upsample from that to higher resolutions.

Here's something to consider. The cells that affect sleep are in such direction that the light coming from overhead has highest effect on it. Oh. And about using phones or computers close to night time, its not necessarily about the light. If you have too much mental or emotional stimulation near sleep time that would definitely harm your ability to sleep. And by mental stimulation, it definitely includes games where being few milliseconds slower means dying in the game.

Air conditioners are sold out every summer when you would really need one.

It's good day to announce this. People are taking all announcements today very seriously. https://www.youtube.com/watch?v=v53EUMNSdQs

Something about all pages being under limit the limit instead of every page being under the limit changes the exact meaning to something that I cannot agree with. Which was the meaning I replied below and then wrote this after realizing you might of meant every when you wrote all.

Lets say you write a daily blog. A single A4 of text contains on average 3000 characters, your posts average slightly above that by being 4000 characters. How long until the text content alone is above the 1MB limit.

https://dictionary.cambridge.org/grammar/british-grammar/all...

The old site looked liked it's exported from org-mode files. New site looks less like that. Probably still done by exporting the same org files in emacs.

The workflow to build that site is completely different from what most web designers are accustomed.

The use youtubeDL for music would probably fit the fair use in Finland. The fair use allows personal copies of legally licensed material but not redistribution of said copies to third parties. It has been used to defend a service where ISP records the specific shows from TV like a VCR and streams them to customers at different time. The ISP:s and representation of copyright holders made a compromise to avoid legal battle, where ISP:s limited the storage duration of said recordings to two years. ISP:s still had service they could sell, and avoid unnecessary legal battle. I'm not a lawyer. So not expert on matter, but just educated local.

That's the Blizzcon 2018 Diablo Immortal reference... Shouldn't take it's literal interpretation as gospel. But overall direction of company seems to be making mobile games for Chinese market.

Sounds to me that they needed to move people from Starcraft team to making mobile version of Starcraft for the Chinese market. They also recognize, that Starcraft is such a phenomena that marketing dollars to get equivalent exposure any other way is going to be a lot more expensive than keeping a small team which keeps it alive.

Just that new and shiny is here doesn't mean the old system became slower. The compilation improvement of 9%, for me it means that most important thing didn't get significantly faster. Unless the Phoronix tests on compiling Apache show better results than AMD:s GCC compilation. (Apache scales badly with threads so, I use that as rough approximation of making change in a file or two and then recompiling only what's affected by the change.)

Using 4k 60hz tv as monitor means CPU gaming performance improvement doesn't matter for 99% of titles as it is limited by either the Screen or GPU.

On the other hand people who have older CPU:s this is really nice increment in addition to previous improvements. A single generation improvement doesn't feel large enough to really matter. Except for professional Gamers with ultra fast GPU:s and monitors, trying to compete for prizes.

If we split the cycle time in two components, one is transistor and one is interconnect. Interconnect delay per length increases as much as the length decreases in each shrink. The transistor delay halves. That halving got transistor delay below the interconnection delay, so doubling that got huge improvements. Also for IPC there is diminishing returns, and easily available improvements got eaten early, and everyone is chasing diminishing returns. The Pentium PRO brought OoO and went from 2 to 3 decode and from 2 to 4 micro-ops/second. Core 2 went from 1 to 2 FP pipelines. Everyone is hitting the same issues, O(n^2) power and die costs for some structures widening things, but smaller fraction of code getting improved by widening. But it isn't all, it also increases latency which is either lower clocks or increased branch missprediction penalty.

Improving CPU:s has become harder simply because all the easier things have been done, and cost of improving one metric often causes worsening another thing slightly.

edit: Just to add, it isn't impossible to improve, but it's just has become harder. And this knowledge was part of reason why until I saw actual benchmarks I wasn't too interested in upgrading. I just found out the one thing I cared had gotten much better during that period.

On package interconnection latency and power is way lower than going off package, AMD also doubled the L3 cache size to compensate for increased memory latency. The issue of putting too much IO in single die is that perimeter of die/package needs to fit all the IO traces on substrate/motherboard, which means more layers and more costs. Everything is just a performance/cost/power trade off. But I would say that off package controllers dealing with multiple CPU chips are probably less viable now than they were before current core count increase.It's because synchronization traffic would require insanely large busses to those controllers, if you wanted to have lots of sockets, and then you would need a lot more pins to those. On the other hand if you had those on package you would get a lot more bandwidth at much lower power and latency, which is what AMD has done.

I thought so also, but decade of small improvements do accumulate to a large number. Something like 20% improvement 4 times would be doubling the IPC, but single threaded compilation is where I got more than that.

The after purchase testing was even better than expected. i7 920 compiled a specific compilation unit in 14 seconds. 3900x in 3.5 seconds. And that was before any tuning, I had done some bios tuning for i7 920, while 3900x I just limited the power to make it quieter. The IPC is more than double, the larger caches are probably the thing that pushes IPC beyond expectations. (I got more cores to improve scaling of multithreaded code.) Both times were recompilations where there was enough ram to have all the files cached and the the folders used where on ssd:s with modern CPU used nvme and older one SATA. Even if the nvme matters that upgrade wouldn't of been possible without getting modern MB. The make system was make, and what I used was LLVM tutorial code in single file that included many LLVM headers. The software wasn't upgraded between those runs, just moved disks from old system to new system and copied the data.

I did look before the purchase if there was IPC improvements that would of made more sense to buy new than buy some Westmere 6 core to upgrade my system. Results made it clear to me that getting any cheap new CPU would of been preferable over wasting time with getting the westmere. The IPC improvements for compilation were way higher than average IPC improvements.

But even sandy bridge was weak enough in the benchmarks that it would of made sense to upgrade from that. Before purchase I just looked from phoronix benchmarks/open benchmarking database a compilation benchmark that had worst CPU scaling for more cores and used that as approximation for single threaded compilation.

My own results were much larger than what I assumed it would of been based on comparing the sandy bridge to modern CPU:s and then multiplying that with clock speed advantage and IPC advantage of sandy bridge vs i7 920.

Oh, when I got i7 920 I decided not to upgrade until I got 8 cores. Then it was AVX-512 happened I knew I must have that so that I could play around optimizing code with it, Intel just could get it's 10nm very soon so that I could get those 8 core AVX-512 parts in reasonable price and power envelope. I just did the math, and realized wasted time because slow CPU would cost me more over next 2 years than upgrading.

I was planning on same thing, until I estimated how much faster modern PC was at single threaded compilation tasks compared to my old one, and how it impacted my productivity. I should of upgraded earlier.

Also the compression algorithm seems to be too computationally intensive for poor server to be hammered by Hacker News crowd.

I think the Babylonians used impaling instead of hanging.

The problem is deciding categories, and what to account for what and duplicate accounting of same operation in different sources and differences in accounting and not everybody doing it the exact same way and lack of transparency for sub-components. Think about making steel for wind turbines, or transporting sand for base of wind turbines. Both actions should be accounted on wind turbines co2 emissions, but also they should be accounted for transportation/steel industry and construction. Everything is interdependent. But lets put things in perspective what is ineffective, in Finland home electricity use is 7% of all energy products used in the country, and still people concentrate on that you should make marginal reductions in your electricity usage to reduce CO2 emissions.

It is my personal opinion that you don't need "*" character ever there are other alternatives available always. And if avoiding that character makes it far safer I'm all for it.

rm -rf /+(?)

If anyone not understanding the joke, + character matches 1 or more instances, and (?) matches any one character. And / is start of entire filesystem. rm is remove. -rf (recursive, never ask verification). So it removes every file that you have rights to remove.

You can also do same thing accidentally by proxy. When my first computing teacher asked entire class to write anything to terminal to show that you had to write specific commands and computers don't understand normal language. Then she told to press enter. I asked her if she was sure. She asked why. I said that I wrote "format c:" . I have never seen teacher walk so fast in a classroom. I think that was last time she used that line in the classroom.

I told you what I did when my first real transformation from unhealthy to healthy happened in my early 30's . Now I'm 41 and I'm having second transformation beyond what happened in the first. Most people really need the first transformation. And one of the driving factor of getting it rolling reasonably well was that I realized how little was truly required to get a lot healthier. There ways to optimize it, but most important thing is to make one simple change and stick with it. Two were real changes I decided were the lifting weights and slowly reducing number of spoons I put sugar in my tea. Everything else was just what I did naturally anyway. The huge point helping me to stick with it was that there were only two simple rules that I had to stick, and both of those were low enough effort for me to stick with.

It wasn't completely without civilization noise, but it was something where I could actually think better when walking outside than when sitting still in front of computer. It was quiet enough that fresh air and walking actually helped me to think.

Taking care of health is interesting optimization problem. Too bad that almost advice in the topic is given by people trying to change fit bodies to professional bodybuilders. Or from crash course to fastest results with diminishing returns of time invested. My personal experience that not exercising takes more time than exercising. Yes, when I didn't exercise, I lost a lot more time of being tired and sick than I spend when I started to exercise.

Here's how I started and lost 10kg in under a year, and more than that in fat but didn't measure fat, but belt holes showed it even when I stagnated in weight. Not diabetic so this is just my personal experience should consider your own body.

1) Lifting adjustable dumbbells at home. Done right before taking shower anyway. So avoiding transition time to exercise and avoiding taking extra shower and drying time can go back to computer right after shower. For beginners maximal muscle growth is 3 exercise sessions per week to have enough recovery between sessions, and single well done set is enough stimulation per muscle, it doesn't grow during exercise but during recover from exercise. I had my exercise sessions on Mondays, Wednesdays and Fridays to be in constantly in a state that if I eat slightly too much it goes to growing muscle instead of fat. Not because maximizing muscle growth, but just to avoid gaining fat at random times. I just had random bodybuilding dumbbell program moves that I thought would hit every muscle and learned those from a book. I didn't have any body weight exercises since moving my weight would of put strain on joints. The time recommendation for that would be more when learning the moves and doing them slowly and checking correct way of doing them and after a while under 30 minutes per session, but keep minimum 10-15 minutes. Now healthy muscles with about a hour per week after initial investment, and not feeling as tired definitely worth doing.

2) Slowly reducing sugar consumption, and my taste buds eventually adjusted to the change.

3) Random eating habits based on if I'm too concentrated on something to miss a meal. If I eat too little I loose fat, if I eat too much I gain muscle.

4) Once I was healthier I spend some of my thinking time just walking outside thinking the problem before returning to computer for actual implementation. Worked since there was quiet forest path near by.

Now my mistake. I upgraded to going gym trying to gain more muscle after my dumbbells became too small. It increased time requirements for each exercise session. After a while I got so busy that "I'll exercise next week", and it stayed that way for a year. And once the habit was broken it was really hard to get back.

I looked Agner's tables, and was curious how Intel fared with it. All numbers are reciprocal throughput. So how many cycles per instruction in throughput. zen 2 has it's gather variants mostly 9 and 6 cycles and one variant with 16. Broadwell has only 6,7 and 5 cycles. Skylake has mostly 4 and 2 and one variant with 5.

Now I was surprised by Agners figures for zen2 LOOP and CALL which both have reciprocal throughput of 2. Being equal to doing with just normal jump instructions.

Skylake on the other hand has 5 or 6 for LOOP and two CALL variants with 3 and one variant with 2.

As a Zen2 owner I'm very disappointed in VPGATHERDD througput, that's so 2013. On the other hand I like the loop and call instruction performance a lot.

Or more likely the 10nm process yield issues are the reason Ice Lake Xeon's are late. One potential is that Ice Lake Xeons are less relevant because it is on market short period before replaced by Sapphire Rapids. Since they got the Tiger Lake Xeons running in lab in June, and it should take them about year to get it to servers. And I doubt that they would have manufacturing issues with Tiger Lake Xeons when they can manufacture Ice Lake Xeons. As for Intel being irrelevant in servers, I doubt that they would go from over 90% market share with on paper inferior products to irrelevant in a year.

"Maybe it's a drill." Said stormtrooper, when Luke and Leia hid and siren was on background.

Finnish translation was originally a physical object drill, and not drill as a practice .

Just took a number which carbon is removed from air-sea-plants natural circulation. The increase in CO2 stops when emissions match the removal of CO2 from circulation as permanent basis. And that figure seems to be about 1% of our emissions. But the key issue is that those two numbers need to match in order to actually stop the climate change permanently which is so far away from every politicians discussion points. All the renewables and storage technologies fail when we start to calculate amount of material required to build them and mining of those resources. Only energy technology so far that could actually do it is nuclear. As for uranium limited reserves they are calculated about 0.1 cent per kwh price limit. It is one of those situations where emotions points towards renewables while numbers are clearly for the nuclear if people could accept tiny risk of having two bananas worth of radiation or even the extreme case of getting so high radiation dose that it would give as high cancer risk as the average alcohol consumption.

The numbers given in article the world wide battery installations are predicted in 2040 to cover about 20 minutes of world electricity consumption assuming it would be equal close to 2016 figures, where majority of world is in relative poverty. It definitely is at least two orders of magnitude too little to stop climate change. People assume the reducing emissions would stop climate change. Nope, emissions are only related to rate of change not absolute value. Reducing emissions by 90% would only slow down the change, it reducing 99% would actually stop it, and it would need to be about 100% to slowly reverse the effects of what has been done by the time humanity gets there.

Its literal translation is.

Satan's cunt-head

Hey that isn't the really impressive, but I did the same thing with DOS configuration files with only using software that came with DOS. I played a game that required 12 megs of RAM and I had 8. I used drivespace to compress the drive and smartdrive to create diskcache. The dos extender of the game automatically swapped to the disk, when there wasn't enough ram. With my set up, the game swapped, then drivespace compressed the data and smartdrive cached the data, and it never hit the disk during a combat anymore and became playable.

The Amdahl's law applies here also, if your code is slow everywhere, making the slowest part faster isn't going to help much, so there needs to be some basic minimal consideration given the performance throughout the project. Or better yet, have well performing default ways of doing things and avoid badly performing ways.

In worst case failing to do that will require write entire project from scratch because basic data structures and structure of code is antithesis for performance and there is not a point that you can optimize. A reasonable default is using vectors over linked lists. A more complex choice is using struct of vectors rather than vector of structs. And neither of those choices are easy to do for data structures that really matter late in the project.