| Page 2 of 2 |
|
|
 Posted: Sun, 1st Dec 2013 15:20 Post subject: |
|
 |
I have no trouble with USB3 on Sandy. Should I have? As far as I know, it's a chipset functionality and not limited by CPU. My Z68 has USB3 and it works fine for both speed (on USB3 devices) and delivering additional power.
|
|
| Back to top |
|
|
|
|
| Posted: Sun, 1st Dec 2013 15:23 Post subject: |
|
|
Your Z68 has USB 3.0 but it's done by a third party controller on the motherboard. The early USB 3.0 controllers have reliability issues and don't have as much bandwidth as they could because they don't support the faster form of data encoding. If it's like mine you've got the option to enable the higher speed method of encapsulation, but even then it's not the /fastest/ and it won't work with every device either. With Haswell it just works. No wonky drivers, and no fucking around with anything. The only time you'll see issues is if you have a shit cable.
And yes, it's a chipset issue but no Sandy CPU can be inserted into a motherboard that has native USB 3.0 support due to the socket change for Haswell. (So really this applies to both Sandy/Ivy) The just of it is that for hassle free USB 3.0 support you *really* want Haswell.
|
|
| Back to top |
|
|
|
|
| Posted: Sun, 1st Dec 2013 15:31 Post subject: |
|
|
Again, I have no hassles with USB3 so I don't see why I want anything else. Perhaps this isn't an "early problematic" controller? I don't get the point of something being "native" anything if it already works - and in my experience, works perfectly - on something else. If the processor doesn't have "native" support then it comes from an expanded chipset, like mine, and ends up being pretty much indistinguishable.
... and no, I'm not being deliberately contrary, haha.
|
|
| Back to top |
|
|
|
|
| Posted: Sun, 1st Dec 2013 15:33 Post subject: |
|
|
Like I said, it was an issue with the early controllers, and if you've got a Z68 chipset you've 100% got an early controller that isn't letting you have the speeds that you should. Not to even mention the fact that the controller is taking up PCI-E lanes to operate which sucks even more when you've got a sandy CPU where you've only got PCI-E 2.0 as it is.
|
|
| Back to top |
|
|
Stige
Posts: 3545
Location: Finland
|
| Posted: Sun, 1st Dec 2013 18:06 Post subject: |
|
|
| Mchart wrote: | | Like I said, it was an issue with the early controllers, and if you've got a Z68 chipset you've 100% got an early controller that isn't letting you have the speeds that you should. Not to even mention the fact that the controller is taking up PCI-E lanes to operate which sucks even more when you've got a sandy CPU where you've only got PCI-E 2.0 as it is. |
I saw zero difference going from 2.0 to 3.0 back when I had HD7950, both on single and dual card setup.
|
|
| Back to top |
|
|
|
|
| Posted: Sun, 1st Dec 2013 18:49 Post subject: |
|
|
| Stige wrote: | | Mchart wrote: | | Like I said, it was an issue with the early controllers, and if you've got a Z68 chipset you've 100% got an early controller that isn't letting you have the speeds that you should. Not to even mention the fact that the controller is taking up PCI-E lanes to operate which sucks even more when you've got a sandy CPU where you've only got PCI-E 2.0 as it is. |
I saw zero difference going from 2.0 to 3.0 back when I had HD7950, both on single and dual card setup. |
It's more about future proofing, and if you use your PCI-E slots for more then just video cards the extra bandwidth of 3.0 helps when you have a limited number of lanes per slot.
|
|
| Back to top |
|
|
|
|
| Posted: Sun, 1st Dec 2013 19:32 Post subject: |
|
|
If it's Renesas/NEC USB 3.0 controller, you can update its firmware and get rid of most bugs.
|
|
| Back to top |
|
|
Stige
Posts: 3545
Location: Finland
|
| Posted: Sun, 1st Dec 2013 19:41 Post subject: |
|
|
| Mchart wrote: | | Stige wrote: | | Mchart wrote: | | Like I said, it was an issue with the early controllers, and if you've got a Z68 chipset you've 100% got an early controller that isn't letting you have the speeds that you should. Not to even mention the fact that the controller is taking up PCI-E lanes to operate which sucks even more when you've got a sandy CPU where you've only got PCI-E 2.0 as it is. |
I saw zero difference going from 2.0 to 3.0 back when I had HD7950, both on single and dual card setup. |
It's more about future proofing, and if you use your PCI-E slots for more then just video cards the extra bandwidth of 3.0 helps when you have a limited number of lanes per slot. |
Don't have any use for the extra slots :l
|
|
| Back to top |
|
|
Frant
King's Bounty
Posts: 24656
Location: Your Mom
|
| Posted: Sun, 1st Dec 2013 22:17 Post subject: |
|
|
| Guy_Incognito wrote: | | If it's Renesas/NEC USB 3.0 controller, you can update its firmware and get rid of most bugs. |
Yeah. I've updated the Renesas/NEC firmware as well as drivers and while I've never had issues (my external HD's using USB3 are just as fast on USB3 as when they were plugged into SATA2 or SATA3) I want to keep my stuff updated. (P67 chipset here)
Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn!
"The sky was the color of a TV tuned to a dead station" - Neuromancer
|
|
| Back to top |
|
|
|
|
|
| Back to top |
|
|
Werelds
Special Little Man
Posts: 15098
Location: 0100111001001100
|
| Posted: Wed, 12th Feb 2014 12:09 Post subject: |
|
|
FX-8320 
That'll keep up with an i5 in games without a single issue (and without Mantle).
Right now though? I'd seriously consider one of the A10's: 7700K or 7850K. TrueAudio comes as a bonus, Mantle works wonders with these fuckers and for 90% of the time the iGPU is good enough. For the other 10% in games you stick an R7/R9 card next to it and crossfire (which with Mantle shows very good results; even adding a 240 or 250 basically doubles performance)
That's assuming Mantle's here to stay though 
|
|
| Back to top |
|
|
|
|
|
| Back to top |
|
|
| Page 2 of 2 |
All times are GMT + 1 Hour |
)
