I 140 Instructions 2006: Fill & Download for Free

You just asked the guy who had this license plate:Bitwise operations have many uses. I see a number of other answers mention “low level code.” Sure, it comes up there. But it comes up in higher level places as well.Low-Level Examples: Focused on hardware interaction and file formats.Manipulating fields in a packed data structure, such as:Hardware registersWire protocolsFile formatsGraphics pixel formatsTranslation table entriesAligning addresses to hardware boundaries such as words or translation pages.Densely packed sets of semaphores, updated by bitwise Atomic Test-and-Set / Atomic Clear instructions. (Note: Links go to recent realizations of these concepts; however, the concepts are older than I am by decades.)Encoding or decoding CPU instructions.OS kernel fault handlers sometimes have to decode instructions to learn more about a fault.Program loaders for dynamically linked executables/libraries usually need to encode to handle relocation entries.Gray code decode for hardware sensorsMid-Level Examples: Focused on bit-level algorithms and mathematics.Error checking and error correcting codesParityHamming codesFinite field arithmetic over power-of-2 fields.CRCs live here.BCH and Reed-Solomon codes also live here.Most cryptographic algorithms.Some numeric formatsExp-GolombMany compression algorithms.Huffman codingShannon-Fano codingBitsliced implementations of algorithms: Each bit within the word represents a different invocation of the algorithm. Within a 32-bit word, you can execute 32 parallel invocations of the algorithm.Convolutional codesEncodingDecodingHigh-Level Examples: Focused on bitwise representation of high level data structures.Set operations: You can represent sets as bitmaps (aka. bit set or bit array), if you can encode the keys as a relatively dense set of integers.Bitwise AND on two bitmaps is set-intersection.Bitwise OR on two bitmaps is set-union.Bitwise AND-NOT on two bitmaps is set-difference. That is, [math]A \wedge \lnot B[/math] is the set of things in [math]A[/math] that are not in [math]B[/math]Bitwise Population Count tells you how many elements are in the set.Collision detection (2-D bitmap graphics): This is actually a special case of set intersection.RTL simulationMachine visionThresholding to a binary imageMorphological image processingBinary image segmentationSpread spectrum communication using LFSRs.Match scoring for large data such as DNA sequences. (e.g. use the Hamming weight of the “not-equal” bit-vector between to sequences to score their similarity.)Arithmetic Tricks / Miscellaneous:The 3-XOR Swap:[math]\begin{eqnarray*} a \leftarrow a \oplus b \\ b \leftarrow a \oplus b \\ a \leftarrow a \oplus b \end{eqnarray*}[/math]Computing unsigned modulo-power-of-2 by AND-ing with [math]2^n - 1[/math];.Rounding down to a power-of-2 by AND-ing with the bitwise-NOT of [math]2^n - 1[/math].These two aren’t quite bitwise, but I’ll include them anyway, because they’re bit manipulation:Dividing an unsigned number by a power-of-2 with a right shift.Multiplying by a power of 2 with a left shift.Vectorization: This intersects with some of the examples I’ve given alreadyComputing predicates for multiple iterations of a vectorized loop.Generating masks for control individual lanes of a vectorized computation.Counting the number of lanes for which a predicate was true or false.The list above is not exhaustive. But, it should give you some idea about the value of bitwise operations.Anecdotes:Recently, I decided to try to optimize this example from Michael Abrash’s Zen of Code Optimization series. The link goes to the version that appeared in his Graphics Programming Black Book.I brought the example into the modern era, and vectorized the heck out of it. The goal of the code is to count words. The problem reduces to: Find and count the transitions from “not-word” to “word” in the input stream. (Or “word” to “not-word”, if you prefer.)In my vectorized implementation, I used vector instructions to classify 32 characters as “in-word” or “not-in-word” in parallel. I then extracted a bitvector from the byte vector, where each bit indicated whether the corresponding character was “in-word” or “not-in-word”. Some bit-shifts and bitwise Booleans yielded a 32-bit bitvector that had a 1 at each transition from not-word to word. One popcount later and I’ve counted all the words that started in that 32-char vector.That code ran several times faster than the scalar original.Another example was a recent homework question I answered: Collect all the strings from the input that have 3 to 10 characters in them, and reject the rest. I ended up writing an optimized, vectorized version that I linked to in this comment. That version makes heavy use of bitwise operations to implement a non-linear FIR. It follows that with rightmost-one detects and bit counts to figure out how may words it’s found and where.Going further back: At previous $DAYJOB, we were trying to win a set-top box socket with our DSP chip. One of the cost adders, though, was decoding the Digital Video Broadcast Common Scrambling Algorithm. The DVB-CSA block cipher wasn’t a problem. The DVB-CSA stream cipher, OTOH, did not yield to software implementation. Our best implementation required 24 cycles/bit, and others on the team that looked at it thought they could maybe get it to 20 cycles/bit.I developed an infrastructure to reduce the 14 S-boxes from ~290 logic operations down to 140. I then implemented a bit-slice implementation that brought the overall cost down to 1 cycle/bit.Now, Wikipedia points out these weaknesses today. In early 2006, no-one publicly had disclosed doing so, [EDIT: Not true. See addendum.] nor had anyone disclosed realizations of the s-boxes that were as compact as mine. A couple years later, libdvbcsa got there.At the time I wrote that code, though, we were so far ahead of the curve that one of the entities involved in approving our solution wasn’t prepared for our highly-performant software solution. (They were in the business of licensing their hardware block to do the same.) Thankfully, we’d architected a secure-algorithm protection solution into our DSPs that made them somewhat, if not entirely comfortable about certifying our design. They certainly didn’t want us to disclose how we’d managed to accelerate the software.AFAICT, I was about 2 years ahead of the public on that one. [EDIT: No I wasn’t. See Addendum.] If anyone can give me better dates on where various aspects of DVB-CSA descrambling were optimized in public codebases, I’d love to know. I did most of my work in March - April 2006.Addendum:It appears FFDeCsa did their bitslice work in 2004. I don’t think I was aware of that work specifically, but I was aware of bitslicing crypto more generally. I know for certain I ended up relying on this site after I started my work. I am pretty sure I had decided to try bitslicing before I discovered that site, though.That site is what gave me my ~290 gate count for the SBoxes. IIRC, attempts to feed the SBoxes there through standard logic optimization tools didn’t bring the count below 270. The FFDeCsa version looks to be 174, which is quite a bit lower. It’s still above the 140 I achieved.

There is little doubt that when the Wehrmacht’s Heavy Tank Battalion 503 first met the IS-2s of Col Tsiganov’s 11th Heavy Tank unit near Tarnapol, the new Soviet tank came as a nasty shock.In May 1944 in Romania, Tiger crews of Großdeutschland near Tirgu Frumos opened fire on IS-2 tanks at long range- and were alarmed to see their previously omnipotent 88 mm rounds bounce off the enemy vehicles [Ref 1]. Ref 2 claims (without reference to the original report, alas) that another in another 1944 incident an IS-2 of 72nd Independent Guards withstood 5 hits from an Elefant tank destroyer armed with the L71 88 mm at very long range.The IS-2 tank’s 122 mm gun could knock out a Tiger frontally at 2000 m. Similarly the IS-2 could defeat the Panther glacis at about 700 m and turret at 2000 m [Ref 3]. General Guderian issued instructions that engagements between even Tigers and the IS-2 should be avoided unless the Germans possessed a tactical advantage.I have compared the tactical value of the Tiger I and IS-2 here.How would ten Tiger tanks compare with 10 IS-2 tanks in combat effectiveness, assuming all tanks are in good working order?In contrast, Ref 4 stated that,using the PaK43 88 mm L71, “all targets T34, KV1 and IS-2 could be destroyed at ranges up to 3000 m”.Based on photographic evidence, combat reports and penetration calculations, the IS-2 could be destroyed at most battle ranges by German weapons, albeit not under all circumstances.The IS-2 had a very strong glacis plate but could be defeated by Wehrmacht Heer 75 mm and 88 mm weapons (and hollow charge weapons such as Panzerfaust) through the turret at all aspects and side/rear. There are graphic accounts of the effects of Panzerfausts on IS-2 tanks in Ref 5Below is an early IS-2 captured by the Wehrmacht, with armour thicknesses stencilled on. So any German gun that could penetrate >100 mm of armour has a reasonable chance against the Soviet beast. Photo credit: Bovington Tank Museum 2449/A1.I have shown elsewhere that even the humble PzKpfw IVH could hold its own with th IS-2 at ranges below 1000 m. see How well could the German PzKpfw IV tank deal with Russian heavy tanks in World War 2?IS-2 (1943 model) penetrated through the upper hull front. An internal explosion has disrupted the turret. What looks like a penetration in the number 2 part of the turret is a pistol-port- you often see holes here in knocked out IS-2s- note the picture below shows the same effect on the other side.Another 1943 model ripped apart.Below is a early model IS-2 destroyed in Berlin -probably by Panzerfaust- it has suffered an internal ammunition explosion- but I suppose the reader has guessed that!There is no doubt that the IS-2 was a tough customer, however, it was not a wonder tank and nor did it give the Wehrmacht the kind of problems that the Matilda II or KV did earlier in the war, insofar that it could be killed by standard tank and A/T guns.In addition, the IS-2 suffered from a number of design weaknesses. Its cast armour was prone to spalling, it had a cramped interior, slow traverse rate and low rate of fire with its large separate loading ammunition. [Ref 6]. Early models armed with the D25 version of the A19 gun had a particularly low rate of fire, the D25T featured an improved breech mechanism. There was no turret basket and a lot of the cartridge cases were stored in boxes on the tank floor. The shells were stored in the turret racking. The loading cycle must have been quite tedious after the first few rounds had been fired. Image below showing ammunition stowage.All models were particularly prone to internal explosions- see photographs. This may be partially a result of separate ammunition but is otherwise surprising as the cartridge cases seem well placed from a safety, if not rate of fire perspective.. see D-10 vs D-25. In fairness, many IS-2s were recovered after being knocked out and repaired. Ref 7 shows that, following and engagement with German tanks, 21 were knocked out but 15 were recovered and repaired subsequently.The IS-2 was one of the best WWII tanks at the job it was mainly designed for - blasting strongpoints. This is reflected in both the selection of its main armament and its ammunition loadout (20 HE, 8 AP). With the dearth (and death!) of Wehrmacht armour at the end of the war, it looks like the Soviets made a good choice! In contrast, the Soviets made over 2200 IS-2s in 1943–44.The problems faced by the Wehrmacht from mid 1944 were legion. By 1945, lack of fuel and the bombing of production and transport facilities had reduced the supply of tanks to a trickle. Almost any decent tank would have given the Wehrmacht lots of problems. See Ref 8 for late war combats between IS-2 and King Tigers in East Prussia.This is why late comer Allied tanks such as Pershing and Comet notched up tank kill numbers that can be counted on the fingers of one hand. see Did US Pershing tanks ever directly engage any German heavy tanks in WWII?ReferencesZaloga S, Kinnear J, Askenov A, Koshavtsev A (1997) Stalin’s Heavy Tanks Concord.Bean T and Fowler B (2002) Russian Tanks of WW2 Ian Allan p140Bird L and Livingston R (2001)World War II Ballistics, Armor and Gunnery Overmatch. p59Hogg I (2002)German Artillery of World War II Greenhill.Bean T and Fowler B (2002) Russian Tanks of WW2 Ian Allan. p139–140Baryatinsky M (2006) The IS Tanks p49–51Baryatinsky M (2006) The IS Tanks p45Higgins DR (2011) King Tier V IS-2 Operation Solstice 1945 Osprey

This answer may contain sensitive images. Click on an image to unblur it.Yes, but I wouldn’t say it’s a belief. Together with the innumerable fascinating and concordant very similar experiences reported here and elsewhere, the conclusion that some kind of “afterlife” certainly exist is inescapable. The alternative would be to consider the thousands of witnesses - from all epochs, cultures, countries, languages, religious or non-religious - telling the almost identical story are liars who have somehow agreed to tell the same lies...So-called “rationalists” and pseudo-scientists try to explain these concurring stories by saying something like “since our brain may still work for some time after dead and since we all are of the same specie, experiencing identical feelings when we die is normal”. But that fails to explain how clinically dead people but brought back to life can accurately report events that happened during their “death” in rooms and places away from their body.I always try to be rational and avoid beliefs, be they positive or negative. These pseudo-rationalists trying to debunk all “afterlife”, NDE and OBE reports are in fact irrational: they hold to their preconceived ideas, to their dogmatic negative beliefs, instead of keeping an open mind, study the reports without prejudice and realize that our knowledge is extremely limited compared to the tremendous complexity of life and universe.Now comes my story.On January 2th 1965, I completely wrecked my father’s brand new MG 1100 in a triple rollover beginning at its top speed - 140 km/h, 87 mph. Luckily, my passengers and me escaped unharmed except a broken tooth, but the trouble was that my father had only a third party insurance for his car. That was already quite destabilizing by itself, but I had another source of deep sorrow.My sister had a pretty and smart 16 y.o. friend named A*. I was 18, this cute girl seemed to like me and we were often together. I started to love her but there was nothing more than friendship from her side and I became despaired, thinking of suicide. One day, I entered a drugstore and asked for parathion, a violent pesticide. I obtained it immediately without any question from the young pharmacist.As I was becoming increasingly frustrated and fed up with my life, a few days latter I left home with the flask of parathion in my pocket and I walked away for several hours till far in the country side. As I walked by a phone booth in some village, I phoned to A* to tell her goodbye. She strongly reprimanded me and told me to come back home, that my parents and my sister were in a terrible state, that henceforth everything will be better and fine for me, they will be more kind with me and so on… She made me realize the pain I would inflict them, so I told her “OK, I’m coming back”.Some time latter, she had come to visit us and I accompanied her by walking to her home, near my parents house. I wanted to give her a present to thank her for having convinced me to come back home instead of committing suicide. But I didn’t know what to choose for her and I thought that giving money would be the best, so that she could buy whatever she liked! I was so stupid, I can’t believe how stupid I was!Of course, it turned into a dispute. I told her “take that money or I throw it away in that water drain-port!”. The night had settled and there was nobody around. We had never hugged, never even had a friendly kiss. 50 years ago, no one did it, except perhaps in exceptional circumstances - or lovers in intimacy. I don’t remember that we ever kissed each other in my family neither. Back to the story: loosing control of myself, I forcefully pulled her in my arms and dragged her against the loading dock of the small dairy products factory we were walking by. She became extremely afraid and tense. “Release me!!”, she said. I wanted to, but I realized that I was already no longer holding her: she was holding me, her hands were clenched on my arms! I said: “but it’s you who’s holding me!” She looked at her hands grasping my arms and she also realized in turn that it was indeed the case. She released her grip and ran away to the entrance of her nearby flat.I came back home. It was late, my parents and my sister were already asleep upstairs. I took the flask of parathion (I had hidden it somewhere, I don’t remember where and why, probably just in case) and drank some. The taste was so atrocious that I was unable to drink more than a little bit. But I knew it was enough anyway. I went to the phone on the wall and called her: “I’m gonna die, A*!” She replied angrily “I don’t care, you spoiled everything, you stupid idiot!”That was it. I was wondering why I wasn’t already dead. Soon, her mother rushed inside our house and loudly called my parents. I brutally tried to silence her but was unsuccessful. Then all my strength left me and I ended up lying on the floor. My family members had come down and they were all standing around me now; after a while two ambulance paramedics arrived also. My mother asked “Where is the packaging of the poison you took?”. I looked at their faces and felt their insuperable pain. I answered “In the garbage can”. As the paramedic were about to load me in the ambulance, I told them “I hope you won’t be able to save me”. It made my sister burst into tears.I must have fallen unconscious since the next I remember is that I very briefly awakened in hospital with medics around the bed and a line in my arm. Then I arrived somewhere else, where there was peace, love, beautiful yellow bright light all over, and some kind of immaterial human beings / souls. It’s difficult for me to describe it all since it was not as neat, or perhaps not as clearly remembered as others have better described. I didn’t see my own body from above neither. What I remember very clearly is that I was instructed not to do it again by an authoritative being / soul with a human body appearance that I could compare to the common representation of Jesus, even if I’ve had no longer any religious beliefs. If I were still believing in Christianity, I would probably ascertain that it was Jesus himself. Or perhaps was he? Or perhaps was it just a representation of something / someone / some universal soul / spirit that we can’t comprehend and that many call God?When I was back in my body again, my state of mind was deeply transformed. I promised my parents and myself that I won’t do it again, in order to keep with the transcendent instruction received.My story doesn't prove anything by itself, but it is consistent with the more complete experiences of others and there's a clear overlap. I’m pleased to share it with you all… this being the first time I tell it to anyone - at least the whole of it. Writing it has caused me deep emotions.The cat knew when my father died.In 2006, my father died in a hospital. We knew there was no hope and that he would very soon pass away. It happened in the middle of the night. My mother was now alone in our house and still sleeping upstairs. At the moment my father died, she was awakened by a very loud meow, so eerie and loud that she though a wild beast had climbed upstairs and was behind her room door. It was the cat.This cat liked my dad very much and was always on his lap. She had never climbed upstairs, it was the first and only time she did.