| Tweetovi |
|
Paolo Bonzini
@fagiolinux
|
2. velj |
|
Don't worry, it's rainy as usual so we will all have a cold soon too.
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
31. sij |
|
It's not like you can launch a new ISA by royal decree, see what happened with IA-64. ARM grew exponentially in the 2000s but it had been around for almost as long as x86. RISC-V is getting there, but it's taken years and many features are still missing. No alternatives for AMD.
|
||
|
|
||
| Paolo Bonzini proslijedio/la je tweet | ||
|
Jen Madriaga
@JenInnovate
|
30. sij |
|
IBM CEO Ginni Rometty is stepping down, Arvind Krishna to take over.
Jim Whitehurst is new President.
cnb.cx/37S9KEt
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
24. sij |
|
Time to bet on a middle seat, cross fingers and hope you get the full row...
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
18. sij |
|
Good luck Laura!
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
11. sij |
|
No link sorry (could also be divide overflow not division by zero, it's the same exception). But if you try Win95 under QEMU it only works with emulation (slower) and not with KVM.
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
10. sij |
|
Nope, modern computers are too fast. Win98 and earlier (possibly ME too) don't boot anymore because of a division by zero.
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
2. sij |
|
Most non-technical people to whom I talk to about my job know that it is a bit weird, so I say colleague and qualify it like "well, not from Red Hat but still..." Within the community I just say "contributor/maintainer" or "the guy/woman/person from company X that does/did Y"
|
||
|
|
||
| Paolo Bonzini proslijedio/la je tweet | ||
|
Julian Stecklina
@blitzclone
|
1. sij |
|
I wrote down my initial stumbling blocks when starting out with @risc_v development: x86.lol/generic/2020/0… I'm probably going to blog more about RISC-V in the future. I should have chosen a better domain name. ;) #riscv #risc_os #osdev
|
||
|
|
||
| Paolo Bonzini proslijedio/la je tweet | ||
|
Software Freedom Conservancy
@conservancy
|
24. pro |
|
Feeling like you want to put some money towards something that matters? Invest in software freedom today -- and get your gift doubled! sfconservancy.org/news/2019/nov/…
|
||
|
|
||
| Paolo Bonzini proslijedio/la je tweet | ||
|
Jens Axboe
@axboe
|
21. pro |
|
Compile times of fio on my laptop, a scientific study:
gcc-5.5: 16.5s
gcc-6.5: 17.7s
gcc-7.4: 18.8s
gcc-8.3: 18.3s
gcc-9.2: 19.7s
Internal benchmark (using null engine):
gcc-5.5: ~8400K IOPS
gcc-9.2: ~8900K IOPS
Compile time increase expected, perf increase nice surprise!
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
18. pro |
|
Yes and the super-secret paging structure cache would probably help with the otherwise large cost of EPT-enabled TLB misses.
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
18. pro |
|
Good point, I suspect all those 4 1GB TLB entries end up caching is the PML4/PDP/PD/PT. :) The worst case then would be a guest that uses 1GiB hugetlbfs and reads random addresses from 32 areas scattered around guest physical address space. 2 MiB hugetlbfs might perform better...
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
17. pro |
|
4GB is still more than what the 2MB DTLB entries can provide though. Also EPT page size can make TLB entries bigger but the size for a combined translation depends anyway on the page size specified by the guest. So you never know but it would be very surprising.
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
15. pro |
|
Sorry but this makes no sense. Sounds very much like moving the goalposts, just a few weeks ago it was more than one hypervisor and there are now three (xmon, KVM and seL4).
|
||
|
|
||
| Paolo Bonzini proslijedio/la je tweet | ||
|
Kristin Russo
@kristinnoeline
|
13. pro |
|
Hello, my name is Kristin and I have never seen a single Star War.
After learning some incredible facts about moisture farming from @jowrotethis I’ve decided to watch every movie over the span of one week beginning with the original trilogy. pic.twitter.com/GDpjhbN6Ir
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
14. pro |
|
Now waiting for freezing of the hypervisor extension...
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
14. pro |
|
Do Risk or chess count?
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
12. pro |
|
I am curious about the Go.
|
||
|
|
||
|
Paolo Bonzini
@fagiolinux
|
7. pro |
|
Unfortunately it ran an old version of Drupal and was hacked. :-(
|
||
|
|
||