digitalmars.D - TechEmpower Web Framework Performance Comparison Round 13 -- vibe.d
- sanjayss (3/3) Nov 17 2016 https://www.techempower.com/benchmarks/#section=data-r13&hw=ph&test=json
- Daniel Kozak via Digitalmars-d (3/5) Nov 17 2016 I guess they have moved from Ubuntu 12.04 to 16.04 which has PIE enabled...
- thedeemon (2/5) Nov 18 2016 Where can I learn more about it?
- Daniel Kozak via Digitalmars-d (2/8) Nov 18 2016 It has been typo, it should not be 16.04 but 16.10.
- qznc (8/21) Nov 18 2016 Why would you run a non-LTS Ubuntu on a server? Anyways, I could
- Daniel Kozak via Digitalmars-d (3/24) Nov 18 2016 It has been a guess, but base on 16.04 which is wrong, so it will be
- thedeemon (3/9) Nov 19 2016 Err.. I mean, what is PIE and what kind of problems with D are
- Chris Wright (8/17) Nov 19 2016 Position-Independent Executable, which is a binary that works with
- Lodovico Giaretta (10/31) Nov 19 2016 As I understand it (I actually use Ubuntu 16.10 and D on it), the
- deadalnix (5/14) Nov 19 2016 DMD emit relocation data that linker can't make sense of, that's
- thedeemon (2/8) Nov 20 2016 Thanks!
- Daniel Kozak via Digitalmars-d (5/11) Nov 17 2016 Btw.
- Daniel Kozak (6/9) Nov 24 2016 So I have investigated it and find out the reason is probably
https://www.techempower.com/benchmarks/#section=data-r13&hw=ph&test=json anyone know why vibe.d is shown as "Did not complete" in all tests?
Nov 17 2016
Dne 17.11.2016 v 18:49 sanjayss via Digitalmars-d napsal(a):https://www.techempower.com/benchmarks/#section=data-r13&hw=ph&test=json anyone know why vibe.d is shown as "Did not complete" in all tests?I guess they have moved from Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.
Nov 17 2016
On Thursday, 17 November 2016 at 18:24:05 UTC, Daniel Kozak wrote:Dne 17.11.2016 v 18:49 sanjayss via Digitalmars-d napsal(a):Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 18 2016
Dne 18.11.2016 v 11:46 thedeemon via Digitalmars-d napsal(a):On Thursday, 17 November 2016 at 18:24:05 UTC, Daniel Kozak wrote:It has been typo, it should not be 16.04 but 16.10.Dne 17.11.2016 v 18:49 sanjayss via Digitalmars-d napsal(a):Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 18 2016
On Friday, 18 November 2016 at 11:02:20 UTC, Daniel Kozak wrote:Dne 18.11.2016 v 11:46 thedeemon via Digitalmars-d napsal(a):Why would you run a non-LTS Ubuntu on a server? Anyways, I could not find anything about the OS on the website. The only info about environment is: Physical hardware environment for Rounds 13 and beyond. Provided by ServerCentral. Dell R910 (4x 10-Core E7-4850 CPUs) application server; Dell R420 (2x 4-Core E5-2406 CPUs) database server; switched 10-gigabit EthernetOn Thursday, 17 November 2016 at 18:24:05 UTC, Daniel Kozak wrote:It has been typo, it should not be 16.04 but 16.10.Dne 17.11.2016 v 18:49 sanjayss via Digitalmars-d napsal(a):Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 18 2016
Daniel Kozak via Digitalmars-d <digitalmars-d puremagic.com> Dne 18.11.2016 v 14:31 qznc via Digitalmars-d napsal(a):On Friday, 18 November 2016 at 11:02:20 UTC, Daniel Kozak wrote:It has been a guess, but base on 16.04 which is wrong, so it will be probably something else.Dne 18.11.2016 v 11:46 thedeemon via Digitalmars-d napsal(a):Why would you run a non-LTS Ubuntu on a server? Anyways, I could not find anything about the OS on the website. The only info about environment is: Physical hardware environment for Rounds 13 and beyond. Provided by ServerCentral. Dell R910 (4x 10-Core E7-4850 CPUs) application server; Dell R420 (2x 4-Core E5-2406 CPUs) database server; switched 10-gigabit EthernetOn Thursday, 17 November 2016 at 18:24:05 UTC, Daniel Kozak wrote:It has been typo, it should not be 16.04 but 16.10.Dne 17.11.2016 v 18:49 sanjayss via Digitalmars-d napsal(a):Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 18 2016
On Friday, 18 November 2016 at 11:02:20 UTC, Daniel Kozak wrote:Err.. I mean, what is PIE and what kind of problems with D are there?It has been typo, it should not be 16.04 but 16.10.Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 19 2016
On Sat, 19 Nov 2016 16:43:40 +0000, thedeemon wrote:On Friday, 18 November 2016 at 11:02:20 UTC, Daniel Kozak wrote:Position-Independent Executable, which is a binary that works with Address Space Layout Randomization. ASLR is a speedbump for attackers -- it puts the stack, heap, and text of your program in random locations on each execution, so an attacker has a lot more work to access globals or specific functions. I don't know what issues D has with them, but I'd guess it has hard-coded locations for global variables or static data or something.Err.. I mean, what is PIE and what kind of problems with D are there?It has been typo, it should not be 16.04 but 16.10.Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 19 2016
On Saturday, 19 November 2016 at 21:26:28 UTC, Chris Wright wrote:On Sat, 19 Nov 2016 16:43:40 +0000, thedeemon wrote:As I understand it (I actually use Ubuntu 16.10 and D on it), the only problem is that to produce PIE executables you need the compiler to emit PIC object code (Position Independent Code). DMD can produce it with the flag -fPIC, but the standard library is compiled without this flag, so the linker rejects it. The current workaround is to dynamically link to the shared version of Phobos, because shared libraries are position independent by default. The other alternative should be to recompile Phobos with -fPIC, but I'm too lazy to investigate this.On Friday, 18 November 2016 at 11:02:20 UTC, Daniel Kozak wrote:Position-Independent Executable, which is a binary that works with Address Space Layout Randomization. ASLR is a speedbump for attackers -- it puts the stack, heap, and text of your program in random locations on each execution, so an attacker has a lot more work to access globals or specific functions. I don't know what issues D has with them, but I'd guess it has hard-coded locations for global variables or static data or something.Err.. I mean, what is PIE and what kind of problems with D are there?It has been typo, it should not be 16.04 but 16.10.Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 19 2016
On Saturday, 19 November 2016 at 16:43:40 UTC, thedeemon wrote:On Friday, 18 November 2016 at 11:02:20 UTC, Daniel Kozak wrote:DMD emit relocation data that linker can't make sense of, that's what happens. So you can't link anything that wasn't compiled with -fPIC, even if you aren't using shared object, and the standard lib doesn't work.Err.. I mean, what is PIE and what kind of problems with D are there?It has been typo, it should not be 16.04 but 16.10.Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.Where can I learn more about it?
Nov 19 2016
On Saturday, 19 November 2016 at 22:19:58 UTC, deadalnix wrote:Thanks!Err.. I mean, what is PIE and what kind of problems with D are there?DMD emit relocation data that linker can't make sense of, that's what happens. So you can't link anything that wasn't compiled with -fPIC, even if you aren't using shared object, and the standard lib doesn't work.
Nov 20 2016
Dne 17.11.2016 v 19:24 Daniel Kozak napsal(a):Dne 17.11.2016 v 18:49 sanjayss via Digitalmars-d napsal(a):Btw. https://www.techempower.com/blog/2016/11/16/framework-benchmarks-round-13/ OK after reading this I am bit skeptical, I do not belive in results anymore :(. Probably is a good time to write another framework benchmarkhttps://www.techempower.com/benchmarks/#section=data-r13&hw=ph&test=json anyone know why vibe.d is shown as "Did not complete" in all tests?I guess they have moved from Ubuntu 12.04 to 16.04 which has PIE enabled now, and there are problems with D on it.
Nov 17 2016
On Thursday, 17 November 2016 at 17:49:09 UTC, sanjayss wrote:https://www.techempower.com/benchmarks/#section=data-r13&hw=ph&test=json anyone know why vibe.d is shown as "Did not complete" in all tests?So I have investigated it and find out the reason is probably some memory leak. It is impossible to run vibed test. Every time it is killed by oom killer or there is a exception about unable to allocate memory. But when I enable VibeManualMemoryManagement version it works ok (2GB+ -> 600MB)
Nov 24 2016