Haswell vs Kaveri

Paran said:
Mobile Richland will be shipped later - in 2014. Richland started shipment in late 2013 and have a look how many notebooks are available. Nothing. Mid-2014 is a very optimistic target for mobile Kaveri. And furthermore there won't be any LGA Broadwells, this is BGA only. If there is a Broadwell for desktop, then these are BGA parts just as Haswell BGA for desktop.
Click to expand...

Nothing but do we know the reasons why? May it be because of business relations, better products from competition, etc?

In general it is more difficult to find AMD products than corresponding ones from Intel and NVIDIA

Oh, and I really hope you are very wrong about the time frame. AMD needs it as soon as possible
 
Paran said:
With Llano it was the same. You can't expect notebooks a couple of weeks after shipment started. This process takes a couple of months until we finally see notebooks in stores. That works for CPU only products in the desktop segment but not for mobile where validating is time-consuming and lots of other hardware parts play a large role. AMD launched mobile Richland in March, paper launch you know.

http://www.pcper.com/reviews/Shows-and-Expos/AMD-CES-2013-Temash-Kabini-and-Kaveri-side-Sea-Islands
Click to expand...

I don't remember anyone else reporting anything like this. I think PC Perspective probably got it wrong here.
 
amd_ces_2013_pressconference_Seite_11.jpg
Click to expand...
AMD-2013-roadmap.jpg
Click to expand...
Kaveri in 2013 is going to be limited to 2-4 core APUs and ultrabook chips, unless something has changed since they released these.
 
I don't expect anything more, I believe Kaveri is a one-of-a-kind chip like Llano and Trinity were. (btw I don't really believe what I quoted about ddr4, ddr4 is not backed up by anything)

Unless there's a big CPU-only Steamroller chip / Opteron die, 8 to 10 cores, but that wouldn't really be a Kaveri.
 
Alexko said:
I don't remember anyone else reporting anything like this. I think PC Perspective probably got it wrong here.
Click to expand...


Because it was an interview with pcper.

The day after the official AMD presentation we were able to sit down with Leslie Sobon for a good hour and really dig into the products we are expecting throughout this next year.
Click to expand...
 
UniversalTruth said:
Nothing but do we know the reasons why? May it be because of business relations, better products from competition, etc?

In general it is more difficult to find AMD products than corresponding ones from Intel and NVIDIA

Oh, and I really hope you are very wrong about the time frame. AMD needs it as soon as possible
Click to expand...
I think we've learned enough from history to safely assume that any delays are invariably AMD's fault. GloFo could probably take some of the blame in certain circumstances, but this certainly isn't one of them.
 
AMD is making people buy a new motherboard every time they update their APU line.

AM2 and AM3 boards lasted over 5 years each. Now they're making people change motherboard every 18 months? This blocks upgrade paths, it isn't nice.

Yes, I do know Intel can be worse though.
 
FM2+ hints at backwards compatibilty though so you could start with a low end APU such as A4-4000 or A6-5300, then swap it later for a Kaveri.
It's how I upgraded from a Sempron LE-1100 to an Athlon II X2 or from a spare old Duron 700 to XP2400+.

If you get the best CPU on a socket from the start you actually rarely have room for meaningful upgrade (barring maybe people who upgraded from Athlon 64 single core to dual core on socket 939, or from dual core AM2 to triple/quad AM3 if the mobo was a nice one). There have often been FSB limitations, or Intel who would change the electric specs without changing the socket's name (socket 775 upgrade hell, three flavors of socket 370).
 
It's an official roadmap, so yes, quite credible. But it's also nearly identical to the one that was released months ago, so I don't understand why you're surprised.
 
It's been quite for quite long that there is no big CPU-only chip made on Steamroller and at the least never a mention of it from sources.
I even remember an AMD slide where Opteron were to be APUs.

We don't know of future plans (Excavator), will AMD make again a big CPU, or an APU with big CPU and small GPU we don't know. Either they've giving up on it or they're waiting for ddr4.
 
As far as I know, Steamroller will power 8-core chips. Just not in 2013.

Same strategy with Piledriver which at first had APU-only incarnations
 
http://www.guru3d.com/articles_pages/core_i7_4770k_review,26.html
Here's another subtlety you should know about, when Intel released Ivy Bridge people noticed that Ivy Bridge processors overheat quite fast once you pass 1.30 Volts on the processor, which has everything to do with the TIM / Intel Heat-spreader used. It has been widely discussed by many of you. With that in mind I was a little surprized to see that with Haswell, Intel decided to ignore the critique and applies exactly the same methodology.
Click to expand...
That is super disappointing, Haswell being a pitiful upgrade over Ivy was expected but to think that the 1 area where Haswell could have redeemed itself was in the overclocking and they couldn't even manage that.
 
Last edited by a moderator:
Wynix said:
http://www.guru3d.com/articles_pages/core_i7_4770k_review,26.htmlThat is super disappointing, Haswell being a pitiful upgrade over Ivy was expected but to think that the 1 area where Haswell could have redeemed itself was in the overclocking and they couldn't even manage that.
Click to expand...

I'm going to update, but I'm still running a Core2Quad that's nearly five years old. Anyone with anything newer like Sandy/Ivy bridge will just be getting incremental improvements that are probably not worth the money.

Bear in mind that Intel don't care about the fraction of a percent of people that will overclock. Intel's sweet-spot is to make more cash by using crappy TIM, and make anyone who wants more buy an E processor and a 2011 motherboard. They don't want to cannibalize higher CPU sales by allowing people to clock cheaper Haswells too far.
 
Wynix said:
http://www.guru3d.com/articles_pages/core_i7_4770k_review,26.htmlThat is super disappointing, Haswell being a pitiful upgrade over Ivy was expected but to think that the 1 area where Haswell could have redeemed itself was in the overclocking and they couldn't even manage that.
Click to expand...
The TIM quality is quite fine. It's just that Intel puts too thick layer of it and that's why above a certain point it goes numb with the heat transfer between the die and the metal cap.
 
DSC said:
http://www.anandtech.com/show/6993/intel-iris-pro-5200-graphics-review-core-i74950hq-tested
Click to expand...

Only around $300 bucks for GT3e huh. I'll take two!

What is the point in this chip? I've read the article twice now to see what I've clearly missed the first time around but the answer is still elusive. Anand appears to believe it's for that group of people who like to pay more for less? No wait - it's all about the power savings, because surely 40 minutes of GT3e gaming is better than 30 mins of some GT2 chip + faster and cheaper discrete?

And still no power benchmarks...
 
Last edited by a moderator:
jimbo75 said:
Only around $300 bucks for GT3e huh. I'll take two!

What is the point in this chip? I've read the article twice now to see what I've clearly missed the first time around but the answer is still elusive. Anand appears to believe it's for that group of people who like to pay more for less? No wait - it's all about the power savings, because surely 40 minutes of GT3e gaming is better than 30 mins of some GT2 chip + faster and cheaper discrete?

And still no power benchmarks...
Click to expand...
Good point about power saving while gaming, not worth the reduced perf over a 650M when plugged in. But discrete isn't only used for gaming. For the people who don't use tools to force integrated only when on battery (discrete kicks on more than you would think, at least on the 15" rMBPs we have at work) or do a lot of unplugged pro work, it could be a decent improvement in battery life, and combined with the supposed cpu power improvements (new low power states, PSR, etc) it could be a big win. I'm speculating a lot here, but I think it's too early to tell until we see actual laptops and real world use case testing.
 
You must log in or register to reply here.
Back
Top