-
Save 15% on ALL SFF Network merch, until Dec 31st! Use code SFF2024 at checkout. Click here!
General Chat Thread
- Thread starter onlyabloke
- Start date
Oh you know, same old same old. Anything awesome I've missed since I've been away?
Looking at building in the Greek A20 but I'm not sure how I feel about Flex ATX PSU.
Looking at building in the Greek A20 but I'm not sure how I feel about Flex ATX PSU.
I meant to put a tongue stuck out face there. Glad you're back and I look forward to your posts again.
Flattery will get you far with me
Kidding.
But I am glad to be back-ish. Keep looking at some of these builds and it makes me want to get mine going.
Just took apart an Xbox 360 and an old PC. Dunno what Im going to do with the old PC crap but the 360...might use the frame and build either a small ITX HTPC style computer or an emulator out of a Raspberry Pi.
Have a drawer full of fans and crap now.
Dont crucify for my storage at the moment. Old as dust components and very limited space.
Have a drawer full of fans and crap now.
Dont crucify for my storage at the moment. Old as dust components and very limited space.
Last edited:
- Jul 19, 2015
- 818
- 1,359
Code:
telnet mapscii.me- Jul 19, 2015
- 818
- 1,359
PCIe 4.0 v1 spec has been finalised and released
https://pcisig.com/sites/default/files/newsroom_attachments/PCIe 4.0 Blog.pdf
https://pcisig.com/sites/default/files/newsroom_attachments/PCIe 4.0 Blog.pdf
So that should mean GPUs can run at 99% of x16 performance on a x2 slot instead of a x4 slot now?
Assuming 4.0 gets adopted for consumers at all; PCIe 5.0 is still scheduled for Q2 2019 spec finalisation, so except for areas where there is an actual demand for more bandwidth (e.g. networking) PCIe 3.0 may stick around until then. 5.0 also still only has PHY backward compatibility as 'targeted', so it may also involve a change in slot type that may be unpopular if it follows a new version only a year or two before.
Personally I'm more hoping that OCuLink might gain more widespread adoption and we might finally be able to ditch the on-board card-edge connector. That would make dramatic layout changes far easier than the current need for ribbon risers.
Personally I'm more hoping that OCuLink might gain more widespread adoption and we might finally be able to ditch the on-board card-edge connector. That would make dramatic layout changes far easier than the current need for ribbon risers.
If PCIe 5.0 requires a new physical connector, I would expect motherboard and add-in card manufacturers to be more willing to jump to PCIe 4.0 first.
Here's hoping that if they do jump on the 5.0 bandwagon from the start, they'll rethink the power infrastructure as well, utilizing 36 or 48V as a base voltage and get rid of the ancient ATX power spec once and for all. Because if you'd need a new motherboard for a new GPU, I could just as well get a new PSU and be set for a more modern standard.
Here's hoping that if they do jump on the 5.0 bandwagon from the start, they'll rethink the power infrastructure as well, utilizing 36 or 48V as a base voltage and get rid of the ancient ATX power spec once and for all. Because if you'd need a new motherboard for a new GPU, I could just as well get a new PSU and be set for a more modern standard.
I'd love to see a move to 48V across the board, but to do that would need a completely clean break from ATX, not just in connector form-factor (as with switching to an all-12V design, where passive adapters would allow the use of existing ATX supplies and existing add-in cards) but a total lack of cross-compatibility and backward-compatibility. A one-two-three punch of new motherboards and CPUs, new GPUs, and a new card interface arriving simultaneously could do it, but ALL of those new parts would need to be a significant enough improvement from their predecessors to offset the huge all-at-once capital outlay and justify not being able to 'get by' with incremental upgrades. That would take a lot of industry co-operation to pull off, and even if Intel have the grunt to push out new from-factors for their own parts, that may just end up like BTX if its reliant on independent manufacturers also taking it up. Even the move from AGP to PCIe had some opportunity for cross-compatible parts (e.g. boards with both slots) which would almost certainly not be possible with a move to 48V (having double sets of VRMs would be both extremely expensive, and likely take up too much board area for GPUs and ITX boards).
Watched Thor. Seems like everyone's started to use disused footage for trailers ._.
Now going back to watching the trailer and I'm going all "did they reshoot that or did they really CG/roto over a rope net?!"
Now going back to watching the trailer and I'm going all "did they reshoot that or did they really CG/roto over a rope net?!"
I like it that way, I severely dislike seeing a trailer that shows most important plot points of a movie. Just as long as the trailer portrays the atmosphere and direction of the film itself.
Example of Too Much Info:
Example of Just Right:
Generally I stick to teasers because many trailers are just movies revealing most of the plot without confirming it. But I want to be kept guessing or atleast be unsure of the plot.
- Jul 19, 2015
- 818
- 1,359
That looks really nice. I’ve made a personal decision to not buy any phone without a headphone jack, but this is very tempting.
Yeah trailers for most movies of the past few years have just been summaries of the movie ._.
there's also the "trailer suggests an extremely different story than the actual movie" one. Maybe I take too much stock in my guesses from the trailer ._.
