1

Topic: intel - still a bug

..So terrible that in fixes in  a kernel slightly  https://www.theregister.co.uk/2018/01/0 … sign_flaw/ https://news.ycombinator.com/item?id=16046636

2

Re: intel - still a bug

Hello, ononim, you wrote: O>. So terrible, what in fixes in  slightly  O> https://www.theregister.co.uk/2018/01/0 … sign_flaw/ O> https://news.ycombinator.com/item?id=16046636 Zalipuhu in Windows thrust a kernel only for ? (Whether to Look in the light of it on AMD as at the processor for home PC?)

3

Re: intel - still a bug

Hello, ononim, you wrote: O>. So terrible that in fixes in  a kernel slightly  O> https://www.theregister.co.uk/2018/01/0 … sign_flaw/ O> https://news.ycombinator.com/item?id=16046636 try {MOV R, byte [R0-addr] CMP R, value JNZ _speculative_predict_false; whether this is taken depends on the branch predictor MOV R1 [cache line N]; speculative fill true _speculative_predict_false:} catch {check if cache line N is filled} it is cool. Went .

4

Re: intel - still a bug

Hello, ononim, you wrote: O>. So terrible that in fixes in  the kernel slightly  O> https://www.theregister.co.uk/2018/01/0 … sign_flaw/ O> https://news.ycombinator.com/item?id=16046636 Looks bad, I throw off actions...

5

Re: intel - still a bug

Hello, ononim, you wrote: Judging by  on IO-bound the progarden will be to 30 % (PostgreSql - 25 %). Interesting as they will roll out a patch. Clouds certainly need to be patched and passed on Ryzen, but with mere mortals what to do not absolutely clearly. I think roll out all from a sin far away, but to gamers and the other expert give the chance to roll away.

6

Re: intel - still a bug

N> Judging by  on IO-bound the progarden will be to 30 % (PostgreSql - 25 %). Interesting as they will roll out a patch. Clouds certainly need to be patched and passed on Ryzen, but with mere mortals what to do not absolutely clearly. I think roll out all from a sin far away, but to gamers and the other expert give the chance to roll away. Under Windows would be the adequate decision to reduce productivity only low-IL processes ( everyones). If such sharing probably generally. Under , accordingly, capability - the general  with a kernel or not.

7

Re: intel - still a bug

Hello, novitk, you wrote: N> I Think roll out all from a sin far away, but to gamers and the other expert give the chance to roll away. Like on gamers a patch and so does not affect, deceleration at a software at which it is a lot of context switchings  a kernel and ring 0. But virtualization, like, can give at all on 30 %, and on 60 %.

8

Re: intel - still a bug

Hello, ononim, you wrote: O>. So terrible, what in fixes in  a kernel slightly  O> https://www.theregister.co.uk/2018/01/0 … sign_flaw/ O> https://news.ycombinator.com/item?id=16046636 the Full particulars from Google, for some reason already uncovered (whether the code already somewhere flowed away that?) https://googleprojectzero.blogspot.ru/2 … -side.html

9

Re: intel - still a bug

It seems that all is absolutely cheerful: https://spectreattack.com/spectre.pdf 4.3 Example Implementation in JavaScript As a proof-of-concept, JavaScript code was written that, when run in the Google Chrome browser, allows JavaScript to read private memory from the process in which it runs (cf. Listing 2). And patches strongly brake CPU for many tasks. Such New Year's  Intel processors on 30 %

10

Re: intel - still a bug

Hello, ononim, you wrote: O>. So terrible that in fixes in  a kernel slightly  O> https://www.theregister.co.uk/2018/01/0 … sign_flaw/ O> https://news.ycombinator.com/item?id=16046636 Lines as already got. It seems that it is time  to pass on AMD?... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

11

Re: intel - still a bug

Hello, CoderMonkey, you wrote: CM> Lines as already got. CM> It seems that it is time  to pass on AMD? Taking into account that complexity of processors approximately equally, is represented improbable that at AMD there is no something similar. Another matter that AMD as a whole is necessary for nobody, and it nobody picks.

12

Re: intel - still a bug

Hello, Sharowarsheg, you wrote: S> Taking into account that complexity of processors approximately equally, is represented improbable that at AMD there is no something similar. Can eat, and can and is not present. And if also is, for the present find.... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

13

Re: intel - still a bug

Hello, CoderMonkey, you wrote: S>> Taking into account that complexity of processors approximately equally, is represented improbable that at AMD there is no something similar. CM> can eat, and can and is not present. And if also is, for the present find. Or already found, only told to nobody.

14

Re: intel - still a bug

Hello, Sharowarsheg, you wrote: S> Taking into account that complexity of processors approximately equally, is represented improbable that at AMD there is no something similar. That's it it (Meltdown) at AMD really is not present: https://lkml.org/lkml/2018/1/3/425 They not "cut off corners" there where cut off . So payment is natural.

15

Re: intel - still a bug

Hello, wildwind, you wrote: S>> Taking into account that complexity of processors approximately equally, is represented improbable that at AMD there is no something similar. W> that's it it (Meltdown) at AMD really is not present: https://lkml.org/lkml/2018/1/3/425 W> They not "cut off corners" there where cut off . So payment is natural. It can and is not present, but as a whole the processor probably is already more difficult, than all entirely 286th together with DOS and Windows 3.0. So appearance of any glitches in equipment is quite natural.

16

Re: intel - still a bug

Hello, novitk, you wrote: N> Judging by  on IO-bound the progarden will be to 30 % (PostgreSql - 25 %). Interesting as they will roll out a patch. Clouds certainly need to be patched and passed on Ryzen, but with mere mortals what to do not absolutely clearly. I think roll out all from a sin far away, but to gamers and the other expert give the chance to roll away. Under a DB to select the separate physical server and to disconnect there this muck, and it is even better to put an old kernel. Well or the processor to change, yes.

17

Re: intel - still a bug

Hello, vsb, you wrote: vsb> Under a DB to select the separate physical server and to disconnect there this muck, and it is even better to put an old kernel. Well or the processor to change, yes. Better  unikernel PostgreSql.

18

Re: intel - still a bug

Hello, Michael7, you wrote: M> Like on gamers a patch and so does not affect, deceleration at a software at which it is a lot of context switchings  a kernel and ring 0. But virtualization, like, can give at all on 30 %, and on 60 %.  gamers it is not assured. On FPS it can and not so, and here a grid/ping should suffer. Anyway to a stench negative PR-a will be if not to give the chance to them much.

19

Re: intel - still a bug

Hello, ononim, you wrote: O>. So terrible that in fixes in  slightly  In bloody  hastily get a kernel  with  and . An epoch euphoria on , "tearing all others CPU as ", seemingly, transit a hot-water bottle.

20

Re: intel - still a bug

Hello, smeeld, you wrote: S> In bloody  hastily get  with  and . An epoch euphoria on , "tearing all others CPU as ", seemingly, transit a hot-water bottle. https://access.redhat.com/security/vuln … eexecution  too covered. About  it is not known.

21

Re: intel - still a bug

Hello, Michael7, you wrote: M> https://access.redhat.com/security/vuln … eexecution  too covered. About  it is not known. Yes it is simple  any... <<RSDN@Home 1.0.0 alpha 5 rev. 0>>

22

Re: intel - still a bug

Hello, Michael7, you wrote: M> Hello, smeeld, you wrote: S>> In bloody  hastily get  with  and . An epoch euphoria on , "tearing all others CPU as ", seemingly, transit a hot-water bottle. M> https://access.redhat.com/security/vuln … eexecution  too covered. About  it is not known. Well, it is new, "glamour" , created explicitly in race for . Early gray-haired  versions 7 and more low - are polished and checked up  as a sea pebble. On  even to stammer does not cost-architecture it is transparent and simple as an axe.

23

Re: intel - still a bug

Hello, smeeld, you wrote: S> In bloody  hastily get  with  and . An epoch euphoria on , "tearing all others CPU as ", seemingly, transit a hot-water bottle. Who and where gets ? Tell Epiki get. At them like the majority of these problems is not present. However, on all does not suffice.

24

Re: intel - still a bug

Hello, Grizzli, you wrote: G> Who and where gets ? Tell Epiki get. At them like the majority of these problems is not present. However, on all does not suffice. Get, there where earlier  this iron, and it heaps have been stored at passage on modern . At us, for example, get. For years on solar oil worked, with  in years.