• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

possible nvidia getting revenge on amd at the kaveri presentation

looks like the nvidia conference is pretty much immediately after the AMD one. wouldn't be surprised if AMD 'accidentally' delayed their conference by about 45 mins or one hour considering what they did at their hawaii event a while ago lol
 
Nvidia are such morons tbh, i know AMD arent totally good guys but Nvidia remind me of shadey guys in dark suits with sungless passing off brown envelopes full of greasy money to buy secret tech and stuff. Bit pathetic really.
 
humbug said:
Tegra is dead and AMD's conference The day after Nvidia, (AMD start tomorrow)

http://www.livestream.com/amdces/vi...0e9c&utm_source=lslibrary&utm_medium=ui-thumb
Click to expand...

Tegra isn't dead, they've just been going the wrong way, if you pump money into failing products you can pay people to put it in products, just ask Intel. More expensive less functional mac air books, just subsidise the chips to the tune of a billion or two and you have ultra books(with no one buying). No one wants Atom's, Intel has multiple times bought products to put it in. Nvidia has done the same and will likely continue to try until the industry actually wants the chip.

Either way AMD did have a conference yesterday, it was just an NDA covered one, not a public one. Public stuff today, from the looks of what's leaked from yesterday of the NDA info, I didn't see much that really needs to be NDA style anyway.
 
Whats wrong with Tegra? I'd accept NVidia over hypes its products but both Tegra 3 and 4 have offered very good performance.
 
Geffen said:
Whats wrong with Tegra? I'd accept NVidia over hypes its products but both Tegra 3 and 4 have offered very good performance.
Click to expand...

Tegra 3 was the reason the original Surface had somewhat pathetic performance (And all the other Windows RT tablets that just flopped, with their crap price and crap performance, although MS is to blame too), it was a SOC that was put into high end devices, but it didn't belong there, while Apple were rolling along with their SOC's that destroyed it, and then Snapdragon came along with their SOC's which were put into lower price points which smashed it all over.

Tegra 3 was not very good, and Tegra 4 isn't all that convincing. The Nokia 2520 in benchmarks I've seen with the Snapdragon 800 does a number on the Tegra 4.
I like my Tegra 4's performance in my Surface 2, but I bet my S800 in my Z1's better.
 
Last edited:
Martini1991 said:
Tegra 3 was the reason the original Surface had somewhat pathetic performance (And all the other Windows RT tablets that just flopped, with their crap price and crap performance, although MS is to blame too), it was a SOC that was put into high end devices, but it didn't belong there, while Apple were rolling along with their SOC's that destroyed it, and then Snapdragon came along with their SOC's which were put into lower price points which smashed it all over.

Tegra 3 was not very good, and Tegra 4 isn't all that convincing. The Nokia 2520 in benchmarks I've seen with the Snapdragon 800 does a number on the Tegra 4.
I like my Tegra 4's performance in my Surface 2, but I bet my S800 in my Z1's better.
Click to expand...

Tegra 3 actually wasn't that bad, but as usual NVidia couldn't come through with decent clock speeds. The Tegra T33 (Surface RT uses T30) enabled 1.6ghz across all cores rather than 1.3ghz and was quite a bit nippier (http://www.anandtech.com/show/6036/asus-transformer-pad-infinity-tf700t-review/3).
It's a shame the Surface RT couldn't have the T33, but I guess NVidia wasn't offering such a good deal on the T30 (NVidia sold the T30 for pennies). You have to consider Tegra 3 was on 40nm too. If they'd managed to shift it to a better process, it could have been a completely different story. But as usual we have TSMC to thank for that.
I had a Nexus 7 for a while and overclocked the pants off the T30L in that. And the performance was pretty good.

Tegra 4 was obviously a complete failure, but Tegra K does look pretty interesting. I'd like to try a Shield though, but shame they don't sell it in the UK :'(

Tegra K will simply be down to how fast NVidia can get it out and how good they can support it. Which they've been terrible at so far...
 
Tegra 3 was a SOC that belongs in 160-200 quid devices, in Android devices such as a Nexus 7 it was stellar. My Dad's running a Nexus 7 and I was using it earlier today, updating it from his 4.2.2 :p.

And it was really quite snappy.

But Surface RT was a 400 quid device, launching at the back end of 2012 (Q4) it wasn't acceptable.
 
Tegra 3's problem was they wanted to be first to a quad and made a HUGE ass chip on 40nm when everyone else knew months later they could do 28nm parts, faster dual cores with faster gpu's and lower cost due to die size, new Arm cores.

They went first, everyone else chose smartly and decided the cost of a new 40nm chip on the brink of 28nm was mental, they were right. THey won a contract due to timing, though had they gone 28nm, Nexus would just have come later, with a 28nm part and Nvidia likely wouldn't have won it. But it's that kind of thing that led to them losing the next Nexus, they were miles behind on 28nm while they were trying to make the follow up Nexus and things so other people won the contracts.

Nvidia seem to be going the same way, yeah, making a 3rd gen 28nm product will obviously give you a huge advantage at 28nm, but they are ploughing money, time and work into 28nm when everyone else is doing it with 20nm chips. Nvidia were late to 20nm and it cost them, and they are repeating history. A epic 28nm chip, is just going to be bigger and worse than a good 20nm chip.

There is some difference, glofo/tsmc are moving to 14/16nm pretty quickly with finfets that should be hugely improved for mobile devices(40% power saving), so missing 20nm wouldn't be a complete disaster IF their next plan includes being on time and competitive at 14/16nm.
 
drunkenmaster said:
There is some difference, glofo/tsmc are moving to 14/16nm pretty quickly with finfets that should be hugely improved for mobile devices(40% power saving), so missing 20nm wouldn't be a complete disaster IF their next plan includes being on time and competitive at 14/16nm.
Click to expand...

They've been in "catch up" mode since Tegra 2 though. We always say "IF", but they never actually pull it off. It's quote thought provoking how a company like this can be so mismanaged :/
 
You must log in or register to reply here.
Back
Top Bottom