AMD on track for launch of Kaveri in February 2014

[SPBHM]

It's barely faster in x86 and loses in gaming. Given the difference in price it's a no-brainer.

source? the Haswell Celeron uses a faster IGP than ivy, should be faster for CPU tasks and gaming,

what is the difference in price!? (the real, the device price)
and if a dual core with no HT can compete with a quad core, the dual core is going to destroy the QC in ST performance, which can affect the user experience I guess.

 

[ShintaiDK]

http://translate.googleusercontent....3.html&usg=ALkJrhhWaq9vg4AwoCZfmG3zrAe-PhN0KA

This new Haswell Celeron champion, the 2955U, is basically no faster than the 1007U that failed to gain any traction vs Kabini.

It's barely faster in x86 and loses in gaming. Given the difference in price it's a no-brainer.

Why do you link an average with graphics performance included. And then come up with the conclusion that Haswell is not much better. Despite its a whole other IGP. The only thing you even got somewhat going for you is gaming.

 

[SiliconWars]

Even without the gaming benchmarks it's only 16% faster overall, and even then it's only because most of the tests are single threaded. These are not fast chips, 16% faster than a low end chip is still low end.

The Haswell only has 200 MHz graphics vs 350 MHz on the Ivy so I wouldn't expect much of a difference if any. We've seen before that there was barely any gains (and some regression) with ULV Haswell's graphics.

 

[SiliconWars]

source? the Haswell Celeron uses a faster IGP than ivy, should be faster for CPU tasks and gaming,

what is the difference in price!? (the real, the device price)
and if a dual core with no HT can compete with a quad core, the dual core is going to destroy the QC in ST performance, which can affect the user experience I guess.

For what it "destroys" the Kabini at single threaded, Kabini "destroys" the Celeron back multi threaded. The biggest lead is 61% for the Celeron and 49% for the A4 so it's more like horses-for-courses.

In the end, nobody is caring about ultimate performance on these and price and graphics are much more important. That's why the ULV Ivy gained zero traction and I'd be very surprised if this Haswell - at even higher price and similar low level performance - changes much. Especially with a turbo Kabini surely on it's way, which is basically all it needs to close the single threaded gap.

 

[Pilum]

So, you assume that Intel's 22nm production prices drop because they upgraded fabs to 14nm production?

Come to think of it, with ongoing process optimizations, production costs may drop slightly. That wasn't what I was talking about, though. Try reading the second sentence in my post.

 

[Erenhardt]

I'm reading Title over and over again and can't find where did you guys got "Intel" from.

 

[Shivansps]

About prices, there is a $250 2955U Chromebook, there is no way a 2955U to cost $132, Dell 11 3000 is at $380 with 2955U with a unknown AMD version at $350, not what ill call a huge price difference.

Just a note, i expect Kabini to be better than a 2955U, but the GT1 has a real chance of caching them up on IGP, and thats my point.

The Gigabyte BRIX Lineup is another example, $300 for a I3-4010, Celeron 2955U is gona be lower than that, where are the Kabinis?
Instead they went directly to Richland with the GB-BXA8-5545.

Even without the gaming benchmarks it's only 16% faster overall, and even then it's only because most of the tests are single threaded. These are not fast chips, 16% faster than a low end chip is still low end.

The Haswell only has 200 MHz graphics vs 350 MHz on the Ivy so I wouldn't expect much of a difference if any. We've seen before that there was barely any gains (and some regression) with ULV Haswell's graphics.

Stop it already, you are just making excuses to make Kabini look better, Haswell and Ivy does have 200 vs 350 mhz igp, those are the idle clocks, both of them have turbo, and we know it uses it because a HD4000 at 350MHZ is not any better than a E-350.

The HD4000 on I3-3217U destroys a A4-5000 by a large margin on games, Kabini does quite OK compared to ULV Pentium and Celeron Ivy Bridges, but Haswell GT1 is a lot faster than that, considering the gain from architecture AND the gain from moving up from 6EU to 10EU.

I'm reading Title over and over again and can't find where did you guys got "Intel" from.

It was about IGP performance on AMD, i say Intel will force AMD hand on having to increase IGP performance no matter what will happen with dgpus, because AMD is forced to keep higher IGP performance than Intel, and cited what is happening right now with Kabini on the low end ultrabook/ultraportables sector, then people started making excuses for AMD.

 

[Erenhardt]

It was about IGP performance on AMD, i say Intel will force AMD hand on having to increase IGP performance no matter what will happen with dgpus, because AMD is forced to keep higher IGP performance than Intel, and cited what is happening right now with Kabini on the low end ultrabook/ultraportables sector, then people started making excuses for AMD.

No, it is about Kaveri release. Graphics forum is this way:
<---

 

[Homeles]

I'm reading Title over and over again and can't find where did you guys got "Intel" from.

If you're looking for context, I can promise you the best place to find it is not in the title.

Wise-guy comments about the direction a thread is taking do not contribute anything to the discussion.

 

[Shivansps]

I did not realise Kaveri has not IGP, sorry.

 

[Erenhardt]

I did not realise Kaveri has not IGP, sorry.

If you want to discuss IGP (integrated Graphics processor) go to Graphics sub-forum.

I was looking for Kaveri detialis, and what I got is 2 pages of intel marketeers comparing 2 year old AMD tech with intel premium parts.

 

[Homeles]

If you want to discuss IGP (integrated Graphics processor) go to Graphics sub-forum.

I was looking for Kaveri detialis, and what I got is 2 pages of intel marketeers comparing 2 year old AMD tech with intel premium parts.

If you have a problem with a derailed thread, report it. Moaning about how things are off topic only contributes further driving things off topic, as you can see with the mess that you have created.

 

[Erenhardt]

If you have a problem with a derailed thread, report it. Moaning about how things are off topic only contributes further driving things off topic, as you can see with the mess that you have created.

Moaning about moaning makes even more mess. Thanks for the suggestion.

 

[Homeles]

Moaning about moaning makes even more mess.

And that absolves you from sin, how? Also, you're making the assumption that I care about threads being off topic -- I don't. I prefer to let them take their course.

 

[AtenRa]

And that absolves you from sin, how? Also, you're making the assumption that I care about threads being off topic -- I don't. I prefer to let them take their course.

Then stop replying to him and only talk about Kaveri, if you have anything to say.

 

[Homeles]

Then stop replying to him and only talk about Kaveri, if you have anything to say.

You missed the part where I said I don't care. I'm here to educate people, regardless of the subject. Today, Erenhardt learned that if he truly cares about a thread staying on topic, he should use the report function.

Not that he really cares, though. He was only here to throw insults at people on the opposing side of the hardware fence.

It's a typical defensive mechanism used by forum-dwellers, including others here. In this mechanism, when a user sees a subject that they don't like (Erenhardt appears to have disdain for Intel, for example), they complain about it in an attempt to suppress information being discussed that isn't relevant to their interests. This behavior is exhibited by both Intel and AMD "fans" alike.

 

[esquared]

Come on people, keep the topic on track here.

If people are derailing the thread, use the report button.


esquared
Anandtech Forum Director

 

[SiliconWars]

nvm, back on topic.

 

[PPB]

OT material, move along.



OnT: Kaveri seems to target 4Ghz, judging by that linkedIn account. So the big IPC boost at the cost of clocks might be discarded already.

 

[frozentundra123456]

OT material, move along.



OnT: Kaveri seems to target 4Ghz, judging by that linkedIn account. So the big IPC boost at the cost of clocks might be discarded already.

Source? Nvm, didn't see it was a linked-in account.

 

[PPB]

Source? Nvm, didn't see it was a linked-in account.

http://www.linkedin.com/in/shardendushekhar


Discredite something without even reading it.... priceless :awe:

 

[Fjodor2001]

http://www.linkedin.com/in/shardendushekhar

Just wondering, where on that linked-in page do you find the info stating that Kaveri should target 4Ghz?

 

[Homeles]

OnT: Kaveri seems to target 4Ghz, judging by that linkedIn account. So the big IPC boost at the cost of clocks might be discarded already.

I believe the primary concern has been that electrical performance would regress as a result of the move from 32nm PDSOI to 28nm bulk, resulting in lower clocks.

Not necessarily directed at you, but something to note as it is commonly misunderstood:
I should also point out that the majority of the IPC increases will be seen when loading the second core in a module, given Steamroller's design. The additional L1D cache and doubled decode hardware is there to keep the execution engines fed when more than one thread is loaded. Bulldozer and Piledriver experience poor scaling when moving from 1T -> 2T within the same module. The main goal of Steamroller is to address this.

What you won't see, based on the information given to us about Steamroller, is a large increase in single threaded performance. You likely won't see 8 decoders being tossed on a single thread, for example, as it would be extremely inefficient. I have no doubts that single threaded performance will improve, but it is important to note that the major changes there will be coming with Excavator.

To reiterate, the largest gains in the Steamroller architecture will be with multi-threaded performance, not single-threaded. At least based off the information we've been given, it will still have relatively weak lightly-threaded performance, however it should achieve the good heavily-threaded performance that AMD intended to gain when they switched from Phenom to Bulldozer.

 

[NTMBK]

Just wondering, where on that linked-in page do you find the info stating that Kaveri should target 4Ghz?

From that linkedin:

Involved in designing synchronous FIFO for Core-Northbridge communication in high performance 4Ghz 28nm microprocessor. Was responsible for RTL and circuit fixes, formal verification, timing fixes and pre-tapeout design checks like fanout, writeability and noise fixes

 

[Ajay]

http://www.linkedin.com/in/shardendushekhar


Discredite something without even reading it.... priceless :awe:

LOL! It is tricky with young engineers who will toss every possible buzz worthy phrase out there to get noticed. Even if it's true, it was a year ago, so it was just a simulation.