|
JawnV6 posted:Actually you're hitting a bunch of elements of x multiple times in a goofy ordering so the data cache is going to wreak havoc on this. Not to mention you're calling an upper bound approximation "proof of the running time" but I guess you're just uh, "trolling" there. It's obvious from context that shrughes means "time complexity" when he says "running time," given that he quoted it in big-O and not microseconds or cycles or something. But I guess you're just uh, "trolling" there. Otto Skorzeny posted:Perhaps I misunderstand you, but on many common architectures integer division on fixed-width types takes a variable amount of time depending on the value of the operands. For instance, on Cortex M3 signed division with 32 bit operands takes ceil(2 * log2(n)) cycles where n is the position of the first leading one in the dividend. Unless you're saying that, for the fixed-width type you always assume the worst case input and thus handwave this logn operation away by calling a 2-to-12 cycle operation a 12-cycle operation in all cases. I never really learned time complexity analysis beyond CS102 level (and definitely haven't done it at this fine-grained a level before, that in any case I don't think most software interviews would care about). In this case of division with a dividend m, division is proportional to log2 n where n is the position of the first leading one in the dividend. Since n is proportional to log2 m, does that mean that 32-bit division on this architecture is O(log log m)?
|
#
?
Feb 10, 2014 20:31
|
|
|
|
| # ? May 22, 2024 05:10 |
|
|
If the number of bits is bounded its O(1).
|
|
#
?
Feb 10, 2014 20:42
|
|
|
JawnV6 posted:Actually you're hitting a bunch of elements of x multiple times in a goofy ordering so the data cache is going to wreak havoc on this. Ask your mom if I care. JawnV6 posted:Not to mention you're calling an upper bound approximation "proof of the running time" but I guess you're just uh, "trolling" there. An asymptotic bound is what the phrase "the running time" generally means. And yes that's a proof that the function runs in O(n^2) time. Yes, that's assuming an abstract machine where dereferencing a pointer or indexing an array doesn't require scanning ~(log n) bits or have speed-of-light limitations implying an average ~(cbrt(n)) random access memory lookup time. shrughes fucked around with this message at 21:11 on Feb 10, 2014 |
|
#
?
Feb 10, 2014 21:09
|
|
|
Good to see you own up to dragging the goalposts around so honestly. I'm pleasantly surprised!
|
|
#
?
Feb 10, 2014 21:17
|
|
|
Talked to the hashrocket guys, the seemed really cool. I think I undersold myself but, whatever. I'd be very happy to do the apprenticeship thing.
|
|
#
?
Feb 10, 2014 21:39
|
|
|
According to consensus, I'm not going to bother getting Programming Interviews Exposed in addition to CTCI, at least not now. Thanks, Cicero, Strong Sauce, and DaVideo!WHERE MY HAT IS AT posted:
Congrats! Are you interviewing anywhere else as well? I'd be interested in hearing how IBM's interview process and eventual offer compares to other companies, since right now it's looking like I will want to leverage my internship experience with IBM to get a position elsewhere in the not-too-distant future. (They just announced more layoffs. IBM's the major tech employer in my region, and I'm thinking escape to NYC after graduation might be a good plan.)
|
|
#
?
Feb 10, 2014 21:43
|
|
Got an offer from IBM!
|
JawnV6 posted:Good to see you own up to dragging the goalposts around so honestly. I'm pleasantly surprised! I don't understand why you're making such a big deal about this. Proofs of algorithm running time and correctness are not some crazy thing that shrughes just made up. Likewise, his assumption that an array look-up or arithmetic operations are constant time is not some radical notion.
|
|
#
?
Feb 11, 2014 00:37
|
|
|
Most complexity analyses don't mention their abstract model. People generally assume a RAM machine where every elemental operation (arithmetic, memory, branching, ...) takes constant time, which is an approximation of the real world that ignores the large constant factor overheads of poor memory access patterns. Word sizes are similarly assumed to be large enough for the numbers involved in the problem, barring algorithms specifically dealing with huge integers. Blowing the cache can greatly increase the constant factor, but the given function is actually Θ(n^2) -- asymptotically bounded on both sides by n.
|
|
#
?
Feb 11, 2014 01:01
|
|
|
bonds0097 posted:I don't understand why you're making such a big deal about this. Proofs of algorithm running time and correctness are not some crazy thing that shrughes just made up. Likewise, his assumption that an array look-up or arithmetic operations are constant time is not some radical notion. It started out as a standard "I know more arch than y'all" jab with an option to open into "oh, so why is your spot on the pedantic spectrum so privileged??" like Otto tried to take it, but the posters eager to smear me coming out of the woodwork to vomit algo 101 basics past my point has been quite a treat!
|
|
#
?
Feb 11, 2014 01:41
|
|
|
Nope, no other interviews scheduled, and I've already accepted IBM's offer. This is just for an internship, and our school's policy is that you're obligated to take the first offer you get (to avoid pissing off potential employers, I guess?).
|
|
#
?
Feb 11, 2014 01:44
|
|
|
JawnV6 posted:It started out as a standard "I know more arch than y'all" jab with an option to open into "oh, so why is your spot on the pedantic spectrum so privileged??" like Otto tried to take it, but the posters eager to smear me coming out of the woodwork to vomit algo 101 basics past my point has been quite a treat! https://www.youtube.com/watch?v=uQl5aYhkF3E
|
|
#
?
Feb 11, 2014 01:45
|
|
|
WHERE MY HAT IS AT posted:Nope, no other interviews scheduled, and I've already accepted IBM's offer. This is just for an internship, and our school's policy is that you're obligated to take the first offer you get (to avoid pissing off potential employers, I guess?).
|
|
#
?
Feb 11, 2014 01:48
|
|
|
JawnV6 posted:Actually you're hitting a bunch of elements of x multiple times in a goofy ordering so the data cache is going to wreak havoc on this. Not to mention you're calling an upper bound approximation "proof of the running time" but I guess you're just uh, "trolling" there. Jawn I know that triggers your autism but usually algorithmic analysis is done in a cache free O(1) ram model cool kids use cache oblivious analysis Malcolm XML fucked around with this message at 01:58 on Feb 11, 2014 |
|
#
?
Feb 11, 2014 01:56
|
|
|
Cicero posted:...what. That is a terrible policy. If that had applied to me, I'd have ended up taking an internship with some random no-name shop in Utah instead of Goldman Sachs, for half the pay (and I wouldn't have been able to go to NYC). I'm quite aware. Hence why I started looking quite early and only apply to the jobs at the tippy top of my list to start.
|
|
#
?
Feb 11, 2014 02:00
|
|
|
WHERE MY HAT IS AT posted:Nope, no other interviews scheduled, and I've already accepted IBM's offer. This is just for an internship, and our school's policy is that you're obligated to take the first offer you get (to avoid pissing off potential employers, I guess?). what kind of nerd school do you go to.
|
|
#
?
Feb 11, 2014 02:11
|
|
|
WHERE MY HAT IS AT posted:Nope, no other interviews scheduled, and I've already accepted IBM's offer. This is just for an internship, and our school's policy is that you're obligated to take the first offer you get (to avoid pissing off potential employers, I guess?). This is a hosed up policy and its hard to believe. How would they even know. Jesus
|
|
#
?
Feb 11, 2014 02:24
|
|
|
an skeleton posted:This is a hosed up policy and its hard to believe. How would they even know. Jesus How would this even be legal? Even if you signed a contract to that effect, I have trouble believing it would be enforceable.
|
|
#
?
Feb 11, 2014 02:56
|
|
|
WHERE MY HAT IS AT posted:Nope, no other interviews scheduled, and I've already accepted IBM's offer. This is just for an internship, and our school's policy is that you're obligated to take the first offer you get (to avoid pissing off potential employers, I guess?). IBM was near the top of your list? Anyway, I'm going to be interning with IBM this summer as well, so if you're in the greater NY area, see you around!
|
|
#
?
Feb 11, 2014 03:12
|
|
|
To clarify a bit on the policy: If you get an offer that comes through our school's co-op department policy states that you're allowed to decline it, however it causes the following things to happen:
So sure, you're "allowed" to decline, but only if you never intend to use any of the co-op department services again (you still have to pay fees for them, of course). I applied and interviewed for the IBM job outside of the school, however they went to the school to make me an offer, which meant I had to take it (I would have taken it anyways in this case, though). Still a lovely policy, but what can you do? edit: IBM wasn't specifically, but the position is at their research labs and they pay quite well (for this area, at least), so whatever. Also I won't see you around because I'm in 
WHERE MY HAT IS AT fucked around with this message at 03:22 on Feb 11, 2014 |
|
#
?
Feb 11, 2014 03:19
|
|
|
He's not the only one... I had a very similar policy at my school (also in . take the first offer given, or else you can't get board access), but in most circumstances, if you were to talk to them, they will usually let it slide. I remember having two offers at the same time, so there was no possible way I can just take the first... At least you are working at IBM an skeleton posted:This is a hosed up policy and its hard to believe. How would they even know. Jesus 
|
|
#
?
Feb 11, 2014 03:27
|
|
|
JawnV6 posted:It started out as a standard "I know more arch than y'all" jab with an option to open into "oh, so why is your spot on the pedantic spectrum so privileged??" like Otto tried to take it, but the posters eager to smear me coming out of the woodwork to vomit algo 101 basics past my point has been quite a treat! Yeah, all offense intended you were just sperging out. Anyone that's been through a basic CS program can tell you that big O approximations are relatively crude. That's not news to anyone here, an n^2 function can run slower than an n^3 function depending on the inputs or the architecture that's running the algorithm. Anybody that's done any assembly will tell you that this is obvious, and I'm guessing that's most of the people here with a formal background. All you've done is gently caress up an architecture 101 concept, and also gently caress up enough that all those people "vomiting algo 101" are more correct than you. One of the basic concepts of big O is that as N increases, algorithmic complexity tends to dominate architectural concerns. If we view a basic x*n^y, then almost all architectural concerns tend to fall into the x, which is the one part that you can ignore. If you know the architecture and you can control the inputs then the arch 101 stuff starts to come into play but since Shrughes specified neither your remark is both rude and idiotic. It sounds like you are simply parroting something you heard in a class or on the internet without really understanding how and why to use the knowledge.
|
|
#
?
Feb 11, 2014 05:40
|
|
|
JawnV6 posted:Good to see you own up to dragging the goalposts around so honestly. I'm pleasantly surprised! 
|
|
#
?
Feb 11, 2014 05:58
|
|
|
|
|
#
?
Feb 11, 2014 09:48
|
|
|
JawnV6 posted:It started out as a standard "I know more arch than y'all" jab with an option to open into "oh, so why is your spot on the pedantic spectrum so privileged??" like Otto tried to take it, but the posters eager to smear me coming out of the woodwork to vomit algo 101 basics past my point has been quite a treat! So what's up with this chip on your shoulder about software people all of a sudden?
|
|
#
?
Feb 11, 2014 09:53
|
|
|
Markov Chain Chomp posted:So what's up with this chip on your shoulder about software people all of a sudden?
|
|
#
?
Feb 11, 2014 10:11
|
|
|
So to derail this terrible argument into one slightly less terrible, should I wear a suit to my next interview?
|
|
#
?
Feb 11, 2014 10:19
|
|
|
Che Delilas posted:So to derail this terrible argument into one slightly less terrible, should I wear a suit to my next interview? If they normally wear khakis and a button-up, yes. If interviewing at Apple, hell no you will be laughed at behind your back.
|
|
#
?
Feb 11, 2014 10:32
|
|
|
wolffenstein posted:If they normally wear khakis and a button-up, yes. If interviewing at Apple, hell no you will be laughed at behind your back. Apple's a sensible enough company about hiring. I interned there and my group hired a guy in a suit. The manager's reaction when he came in was to say, "awww you wore a suit, that's cute. I told you you didn't have to." Then an engineer came in in bare feet and cut-offs and they all grilled the interviewee in front of a whiteboard. He clearly knew his stuff, so they hired him and nobody mentioned the suit-wearing. When he started work, he dressed better than the rest of the team (khakis and button-ups) and nobody cared.
|
|
#
?
Feb 11, 2014 13:24
|
|
|
Mniot posted:Apple's a sensible enough company about hiring. I interned there and my group hired a guy in a suit. The manager's reaction when he came in was to say, "awww you wore a suit, that's cute. I told you you didn't have to." Then an engineer came in in bare feet and cut-offs and they all grilled the interviewee in front of a whiteboard. He clearly knew his stuff, so they hired him and nobody mentioned the suit-wearing. When he started work, he dressed better than the rest of the team (khakis and button-ups) and nobody cared. Totally off-topic I suppose, but I don't think I'd want to work next to a programmer in bare feet (sounds smelly). "No shoes, no shirt, no coding" seems like a reasonable policy. Admittedly, I just wear slippers all day.
|
|
#
?
Feb 11, 2014 14:29
|
|
|
NovemberMike posted:Yeah, all offense intended you were just sperging out. Anyone that's been through a basic CS program can tell you that big O approximations are relatively crude. That's not news to anyone here, an n^2 function can run slower than an n^3 function depending on the inputs or the architecture that's running the algorithm. It was poorly worded but this was the point I was initially trying to make. Sorry.
|
|
#
?
Feb 11, 2014 14:37
|
|
|
NovemberMike posted:Yeah, all offense intended you were just sperging out. RICHUNCLEPENNYBAGS posted:It was poorly worded but this was the point I was initially trying to make. Sorry. Has everyone taken their shot? Anyone else feel like vomiting up an algo 101 post ostensibly aimed at me? Markov Chain Chomp posted:So what's up with this chip on your shoulder about software people all of a sudden?
|
|
#
?
Feb 11, 2014 16:15
|
|
|
heh
|
|
#
?
Feb 11, 2014 16:41
|
|
|
bonds0097 posted:Totally off-topic I suppose, but I don't think I'd want to work next to a programmer in bare feet (sounds smelly). "No shoes, no shirt, no coding" seems like a reasonable policy. Admittedly, I just wear slippers all day.
|
|
#
?
Feb 11, 2014 16:53
|
|
|
Cicero posted:At Amazon everyone always had shoes, but at Google a few more senior engineers I've seen have been barefoot. It looks...odd. I dunno, maybe I'll get used to it. Not that it would change my decision to work there, but man that's weird. Also I start one week from today and the Dunning-Kurger or whatever is setting in.
|
|
#
?
Feb 11, 2014 17:18
|
|
|
Cicero posted:At Amazon everyone always had shoes, but at Google a few more senior engineers I've seen have been barefoot. It looks...odd. I dunno, maybe I'll get used to it. I haven't actually gone barefoot at work, but I get it. Shoes are feet prisons.
|
|
#
?
Feb 11, 2014 18:00
|
|
|
Edly posted:I haven't actually gone barefoot at work, but I get it. Shoes are feet prisons. This. I'd never really considered but if it was an option then I think I would be one of those people
|
|
#
?
Feb 11, 2014 18:25
|
|
|
I have that skype call with Hashrocket in about an hour. Wish me luck! I'm trying to think of questions to ask afterwards, and I'm wondering if it's too presumptuous to ask details about their paired programming interview before I'm asked to do it. I already know that they do that, but I'm not entirely sure if apprentices do the same. Is it okay to ask?
|
|
#
?
Feb 11, 2014 19:57
|
|
|
Pollyanna posted:I have that skype call with Hashrocket in about an hour. Wish me luck! I talked to them yesterday and it's definitely fine to ask. The apprentice thing is very unusual so I think asking questions about is a good idea. It's a very casual interview -- just a screening. It will be much better than my stupid questions. They're very nice, don't stress and you'll be fine. (I wasn't because I stress and then my mind goes blank and I can't remember anything I've ever done).
|
|
#
?
Feb 11, 2014 19:59
|
|
|
USSMICHELLEBACHMAN posted:I talked to them yesterday and it's definitely fine to ask. The apprentice thing is very unusual so I think asking questions about is a good idea. It's a very casual interview -- just a screening. It will be much better than my stupid questions. Yeah, I was coming up with a bunch of questions related to it before now. Still got about 45 minutes! I'm glad to hear that they're nice. I'm still kinda worried about the entire thing, myself, but I'll judge that after I'm done.
|
|
#
?
Feb 11, 2014 20:18
|
|
|
|
| # ? May 22, 2024 05:10 |
|
|
Pollyanna posted:Yeah, I was coming up with a bunch of questions related to it before now. Still got about 45 minutes! Good luck, let me know how it goes.
|
|
#
?
Feb 11, 2014 20:31
|
|