Go to advanced search

by bstrobl
Wed Jun 29, 2016 3:43 pm
Forum: General discussion
Topic: 64bit vs 32bit benchmark. Eben was right
Replies: 65
Views: 21891

Re: 64bit vs 32bit benchmark. Eben was right

Heater wrote:It seems to be a rule that every programmer thinks every other programmers code is garbage.
You don't even need to go that far, my code looks terrible to me 3 months later when a bugfix is required.
by bstrobl
Wed Jun 29, 2016 1:48 pm
Forum: General discussion
Topic: 64bit vs 32bit benchmark. Eben was right
Replies: 65
Views: 21891

Re: 64bit vs 32bit benchmark. Eben was right

It's implied by their memory requirements. And the fact that all large pieces of software are badly written to some extent. Linux kernel may be an exception. To be fair, I actually blame most website designers for overbloating their pages with Ads, Tracking, fancy animation and other unnecessary to...
by bstrobl
Sun Jun 26, 2016 12:15 pm
Forum: General discussion
Topic: 64bit vs 32bit benchmark. Eben was right
Replies: 65
Views: 21891

Re: 64bit vs 32bit benchmark. Eben was right

2. Although pointers will be 64bit, most data they point to isn't, and the number of pointers in code is a lot lees than overall data. So the memory hit is fairly insignificant. It may be more than that. From here . While the increased memory usage may not be critical it may make your instruction a...
by bstrobl
Sat Jun 25, 2016 10:26 am
Forum: General discussion
Topic: 64bit vs 32bit benchmark. Eben was right
Replies: 65
Views: 21891

Re: 64bit vs 32bit benchmark. Eben was right

While there are definite advantages to a 64 bit system the problem is currently that, unless relative addressing is used, pretty much all pointers are going to be 64 bit and hence there will be more cache pressure and higher RAM use even with 32 bit instructions. This is problematic on a system with...
by bstrobl
Fri Jun 24, 2016 8:05 pm
Forum: General discussion
Topic: 64bit vs 32bit benchmark. Eben was right
Replies: 65
Views: 21891

Re: 64bit vs 32bit benchmark. Eben was right

The issue with 64 bit is not performance per se NOT EVEN. ... the issue is entirely computational power (large numbers without big num libraries) and memory capacity. Just to let you folks know (if you're not aware by now already) I'm running a PineA64 as my primary desk PC for the past three weeks...
by bstrobl
Thu Jun 23, 2016 9:39 pm
Forum: General discussion
Topic: 64bit vs 32bit benchmark. Eben was right
Replies: 65
Views: 21891

Re: 64bit vs 32bit benchmark. Eben was right

There should not be such a massive discrepancy between bitness so I think comparing the same version of linux distro makes more sense (i.e. 32 bit Arch to 64 bit Arch). Regardless, due to the 1 GiB RAM limit and overheating issues with the Pi 3, I think the foundation will release a Pi 4 with more e...
by bstrobl
Tue May 31, 2016 7:30 pm
Forum: Advanced users
Topic: Linux kernel is now 4.4
Replies: 114
Views: 69827

Re: Linux kernel is now 4.4

Good, great choice of default action for those of us who want to embrace the twenty year old improvements to IP networks that should have been rolled out Worldwide at least ten years ago. IPv4 needs to die. Going to have to agree here, the default should be enabled unless explicitly disabled. IPv4 ...
by bstrobl
Mon May 23, 2016 3:43 pm
Forum: Bare metal, Assembly language
Topic: Entering aarch64 execution state
Replies: 171
Views: 81656

Re: Entering aarch64 execution state

Actually just went back and checked the documentation on the Cortex A32, and am now beginning to suspect you guys might use it for your next chip :lol: -10% more efficient than A35 -25% more efficient than A7 -A35 is 25% more efficient than A53 -Suggested use of A32 is in wearables and SBCs -VC4 is ...
by bstrobl
Mon May 23, 2016 2:57 pm
Forum: Bare metal, Assembly language
Topic: Entering aarch64 execution state
Replies: 171
Views: 81656

Re: Entering aarch64 execution state

Seems like the Cortex A32 cores would have made more sense for the Pi 3, since ditching the 64 bit instructions alone gives a 10% improvement in efficiency for normal 32 bit operations. Seeing as the A32 processor was only announced a couple of months ago (about the same time as Pi3 was released) i...
by bstrobl
Mon May 23, 2016 2:00 pm
Forum: Bare metal, Assembly language
Topic: Entering aarch64 execution state
Replies: 171
Views: 81656

Re: Entering aarch64 execution state

Seems like the Cortex A32 cores would have made more sense for the Pi 3, since ditching the 64 bit instructions alone gives a 10% improvement in efficiency for normal 32 bit operations.
by bstrobl
Sat May 21, 2016 8:53 am
Forum: Networking and servers
Topic: Idle temperature for a Pi web server
Replies: 6
Views: 1242

Re: Idle temperature for a Pi web server

This is my Raspberry Pi 2 hosted in a datacenter in the Netherlands: [email protected] ~ $ uprecords # Uptime | System Boot up ----------------------------+--------------------------------------------------- 1 135 days, 03:30:18 | Linux 3.18.11-v7+ Fri May 8 10:46:16 2015 2 71 days, 20:41:00 | Linux ...
by bstrobl
Sat Mar 26, 2016 1:43 pm
Forum: General discussion
Topic: How reliable is the Raspberry Pi 3
Replies: 19
Views: 10696

Re: How reliable is the Raspberry Pi 3

If it's only for playing video, get a pi b+. Cheaper, less power hungry and well tested at this stage. Right now the Pi 3 seems to suffer in certain cases with overheating issues, so being able to use a board with lower energy requirements will also mean less waste heat, which when installed behind ...
by bstrobl
Thu Mar 24, 2016 11:22 am
Forum: Advanced users
Topic: store logs on usb flash drive for 24/7 operation
Replies: 8
Views: 2487

Re: store logs on usb flash drive for 24/7 operation

@bstrobl: Thank you for your advice. I took a look into your mentioned video. I think the simplest solution is to use a SD card with wear leveling. I found out, that the memory controller needs free space, so that it can distribute write access equally over free memory sectors. But how can I make s...
by bstrobl
Sun Mar 20, 2016 11:29 am
Forum: Advanced users
Topic: store logs on usb flash drive for 24/7 operation
Replies: 8
Views: 2487

Re: store logs on usb flash drive for 24/7 operation

Pretty much all flash memory has some form of bad block management, since certain parts are heavily written to (e.g. FAT allocation table). Due to the manufacturing process some cells may also be useless after a couple of writes, hence most good SD cards need to do this otherwise they would be long ...
by bstrobl
Sat Mar 19, 2016 12:48 pm
Forum: Advanced users
Topic: store logs on usb flash drive for 24/7 operation
Replies: 8
Views: 2487

Re: store logs on usb flash drive for 24/7 operation

Get a good micro SD card like the SanDisk High Endurance cards. Leave the logs on the micro SD card and make sure you have a good power supply (very important). Most cheap USB flash drives have no wear levelling that I would be aware of, since they are meant to be used intermittently. The SanDisk Hi...
by bstrobl
Thu Mar 03, 2016 7:34 pm
Forum: General discussion
Topic: Future gpu upgrades
Replies: 47
Views: 11884

Re: Future gpu upgrades

I am pretty sure the Foundation will stick with the VC4, even if it costs a pile of money to extend it or whatever. Reason being, they have hired Eric Anholt to do the drivers. Imagine the conversation if the 2838 popped out next year with Mali or something: Foundation: Well, we have got this new 28...
by bstrobl
Wed Mar 02, 2016 8:52 pm
Forum: General discussion
Topic: Future gpu upgrades
Replies: 47
Views: 11884

Re: Future gpu upgrades

There are a number of options, all expensive (going to 28nm would be very very expensive). Adding gig ethernet, improving the memory controller, adding USB 3, adding quads (not necessarily an improvement interestingly) etc. All possible, al expensive. I am curious though, how on earth did you guys ...
by bstrobl
Wed Mar 02, 2016 5:57 pm
Forum: General discussion
Topic: Future gpu upgrades
Replies: 47
Views: 11884

Re: Future gpu upgrades

The problem is the VC4 has been pretty much extended to its maximum capability. One of the big issues that still remain is the 1GB RAM barrier. My guess: The Foundation will invest in either getting that VC5 core up and running or hacking up the VC4 architecture to add additional QPUs along with ext...
by bstrobl
Mon Feb 29, 2016 4:31 pm
Forum: General discussion
Topic: Why didn't we go with Mali?
Replies: 157
Views: 26352

Re: Why didn't we go with Mali?

Would replacing the Cortex A7 with a Cortex A53 and extending memory support to 4GB+ be feasible while keeping the VC4? I remember ARM mentioning the fact that the Cortex A53 is meant to be a drop-in 64 bit replacement for the A7. Edit: Ah here it is - http://www.anandtech.com/show/7591/answered-by...
by bstrobl
Mon Feb 29, 2016 4:09 pm
Forum: General discussion
Topic: The Raspberry Pi 3 Model B Q&A thread
Replies: 457
Views: 212903

Re: The Raspberry Pi 3 Model B Q&A thread

Congrats on the new release and birthday :) Honestly a bit surprised about the BCM2837, given the fact that its only a year after the BCM2836. The Foundation seems to have quite a bit of cash lying around to afford these SoC improvements :lol:. A bit concerned about Eben saying that the B+ might out...
by bstrobl
Mon Feb 22, 2016 6:19 pm
Forum: General discussion
Topic: Best MicroSD card to buy?
Replies: 2
Views: 1059

Re: Best MicroSD card to buy?

You might want to give the Sandisk High Endurance Micro SD cards a try. The 32 GB version is designed to sustain a theoretical 50 TB of writes, while the 64 GB version is meant for 100 TB. Speed is sacrificed for writes in both cases (not an issue since the Pi maxes out pretty quickly on those).
by bstrobl
Tue Feb 09, 2016 10:06 am
Forum: General discussion
Topic: New GPU drivers in raspbian???
Replies: 48
Views: 6413

Re: New GPU drivers in raspbian???

GTR2Fan wrote:I don't know. I'd rather wait until tomorrow for an official statement than spend the intervening time needlessly speculating. ;)
My whole life is based on needless speculation :lol: . Anyone know the new kernel version (4.1.x)?
by bstrobl
Tue Feb 09, 2016 9:54 am
Forum: General discussion
Topic: New GPU drivers in raspbian???
Replies: 48
Views: 6413

Re: New GPU drivers in raspbian???

ShiftPlusOne wrote:It was backported to 4.1
Why go through all the trouble of backporting it to an old Kernel? Is it easier to test for regressions or is the newer 4.4 kernel going to take a lot longer before it becomes available?
by bstrobl
Sat Jan 30, 2016 8:12 pm
Forum: General discussion
Topic: Impact of New Linux Kernel
Replies: 9
Views: 2387

Re: Impact of New Linux Kernel

And yes, most of the stuff in the blob is covered under patent - either from companies licencing the codec(s) or because the interfaces are Broadcom proprietary. A FOSS blob would probably have none of these features available. Thanks for the info. I take it that because of a loss of all those feat...

Go to advanced search