AMD hints at high-performance Zen x86 architecture

[Aquinus]

Ehhhh wrong. For slow vs slower comparisons, the 5350 wipes the floor with the J1900 in most areas. And also has a comphy / minimum 400mhz of overclocking headroom without touching the voltage - on the stock cooler and is a socketed platform. Also the J1900 does not even support sata 6Gb/s. In this very small segment, AMD wins. Definitely not a warning shot... maybe a nerf dart at best while yelling... JUST KIDDING!

Oh.. and definitely not 1/2 the power despite the ratings... A small difference for a large (if you can call it that) performance increase.
http://www.tomshardware.com/reviews/athlon-5350-am1-platform-review,3801-9.html

(oh dear lordy- someone was wrong on the internets and I just felt I had to correct them - shame on me lol)

Ah I remember the days when Intel had nothing on AMD and the barton cores were pissing all over the polished shoes of Intel execs. But that has been many moons. I still support AMD even (buy and build) though I have many intel based PC's. All one can do is hope - for another run!

While I agree the numbers are closer and at that point really doesn't matter. I think you need to factor in cost. You can get a J1900 embedded board for 70 USD, you're practically paying that for the 5350 alone, plus add the cost for the motherboard to go along with it. For the cost, the J1900 isn't a bad option. That's really my point.

Also with respect to power, that's overall draw in that review. I would call the PSU used into question. They're using an 850-watt PSU to test these boards which are known to run well under 60 watts. Not to rail on Tom's, but a PSU drawing very little current on a huge PSU is going to be highly inefficient and if the load numbers are from draw off the wall, I think they're probably not really useful as any changes in usage would be poorly reflected by a meter on the wall. So I have reservations about the power consumption figures here.

Testing these with a 100-watt PSU or a pico-PSU would have given more realistic results IMHO.

 

[eidairaman1]

I hope Amd learned its lesson after Phenom 1 and Bulldozer. They need a star like 462 And 939/940

 

[Aquinus]

I hope Amd learned its lesson after Phenom 1 and Bulldozer. They need a star like 462 And 939/940

I would like to see a performance variant of AMD's Jaguar cores. I think they're on to something there, if you keep the CPU core smaller and more simple (with respect to the pipeline) branch miss-predictions won't just hurt you less, they'll occur less often too because you can optimize for a smaller pipeline like Intel did. The single biggest flaw with the Pentium 4 was the huge size of the pipeline under Netburst and it's the problem that's plaguing AMD now with their module design (with respect to single-threaded performance).

Time will tell though. All we're doing is trying to predict the future and I'm sure we'll all be wrong in the end.

 

[RealNeil]

I hope that AMD comes out with something competitive. This is only good for us consumers when they do. Without AMD being around, Intel can charge whatever they want for their parts. And we know that they'll hammer us good if they can get away with it,...........for this reason, I wish AMD luck.

Most of my gear is Intel, but I have a few AMD systems too. Nothing wrong with the FX-8350 or the FX-6300 at all.

 

[ThE_MaD_ShOt]

Well at least Amd isn't killing off the desktop performance CPU line like so many though they where. Happy

 

[Frick]

TBH I'm more interested in how their line of APU's will shape up before then. Not that it's bad now.

 

[Nokiron]

Then why does W7 for instance separate 32 bit programs to Program Files (x86), and 64 bit ones to Program Files? It's kinda misleading. Then again, that's MS.

Its not really necessary, just convenient. It has its uses.

 

[d1nky]

This Is FXing Serious


In their own words...... Ive never see such (looking for a word to match) marketing!

 

[Jhelms]

While I agree the numbers are closer and at that point really doesn't matter. I think you need to factor in cost. You can get a J1900 embedded board for 70 USD, you're practically paying that for the 5350 alone, plus add the cost for the motherboard to go along with it. For the cost, the J1900 isn't a bad option. That's really my point.

Also with respect to power, that's overall draw in that review. I would call the PSU used into question. They're using an 850-watt PSU to test these boards which are known to run well under 60 watts. Not to rail on Tom's, but a PSU drawing very little current on a huge PSU is going to be highly inefficient and if the load numbers are from draw off the wall, I think they're probably not really useful as any changes in usage would be poorly reflected by a meter on the wall. So I have reservations about the power consumption figures here.

Testing these with a 100-watt PSU or a pico-PSU would have given more realistic results IMHO.

Correct, their numbers are in fact about 10W high across the board with both setups in that review - so good point - but they are still directly comparable as the same supply was used. Both are power sippers, almost silly to argue over a 10-15w difference between them (at the most). One of my own 5350 systems is averaging 25W since the day it was built. Very efficient!

http://www.techspot.com/review/806-amd-kabini-vs-intel-bay-trail-d/page8.html

Both are workable for most office / work situations but doing anything with 3D, the 5350 is the clear winner and a better value. Also with the no issue / virtual guaranteed 400mhz overclock minimum - it does put the slow 5350 even further ahead on all fronts (it was already ahead to start with). I cannot with good intentions recommend a J1900 based system over an AM1 5350 based system is all. Hard to recommend either actually unless someone has a specific need or want as I did. And for my purposes / business - the 5350 rocks!

 

[BiggieShady]

AMD cache design is ridiculous - enormous L2/L3 cache with high latency. It takes twice as long to access L2/L3 cache compared to Intel.
It's ironic to have a huge chip with enormous amount of slow cache that makes it expensive to produce and hinders the performance at the same time forcing them to price them just above the producing cost.
What they need to do is: fast cache in lower amount instead of slow cache in huge amounts, less cache sharing between cores/modules.
...
On the other hand, it's entirely possible that they were unable to produce cache on die that would work with low latencies at desired clock.

 

[Dent1]

I hope that AMD comes out with something competitive. This is only good for us consumers when they do. Without AMD being around, Intel can charge whatever they want for their parts. And we know that they'll hammer us good if they can get away with it,...........for this reason, I wish AMD luck.

Most of my gear is Intel, but I have a few AMD systems too. Nothing wrong with the FX-8350 or the FX-6300 at all.

Whilst what you say is generally true as far as being good for the consumers.

People have to realise that AMD being around period is good for consumers whether they have more competitive performing processors or not.

We've reached a point where most mainstream processors can fulfill the needs of the majority consumer and enthusiast. This has forced Intel to keep their prices reasonable because consumers are becoming savvy enough to understand that a cheap FX 4xxx will push 60 FPS for half the cost of the i7.

I want to see more competitive performing processors too, but for my own selfish and narcissistic needs.

 

[eidairaman1]

Well in the event AMD does pull a good chip out and 1 or 2 units support AM3+ ill have to see reviews to see if they are like a 3.8L Camaro or a Toyota Prius.

 

[RealNeil]

People have to realise that AMD being around period is good for consumers whether they have more competitive performing processors or not.

I agree with you, but if they do well selling, they'll have more resources for R&D.

 

[Dent1]

I agree with you, but if they do well selling, they'll have more resources for R&D.

They can do well selling without a competitive performing processor. Even Intel will make more money from atoms, i3s and i5 than i7s.

The company with the biggest market share wins. CPUs for tablets, laptops, mobile phones, consoles, servers etc. concentrating on just high end desktop won't help them.

Because the console market AMD's market share grew 2.6% this year, Intel has fell by 2.4%..

 

[Super XP]

AMD's Zen Architecture - Built from the Ground Up.
http://en.wikipedia.org/wiki/AMD_Zen

14nm AMD Zen CPU Will Have DDR4 and Simultaneous Multithreading
http://news.softpedia.com/news/14nm...-and-Simultaneous-Multithreading-471401.shtml

Lets not forget AMD has Jim Keller, most likely the greatest CPU architect in the world. Based on Intel, Apple, Microsoft, Nvidia and IBM.

Zen is being designed to not only compete with Intel's future CPU's, its being designed to overtake them. Never underestimate a company that's been backed into a corner for so long.
After his work on K8, Keller left AMD in 1999. That is when the company started to go down hill.

He's been working on this ZEN project for some time now. I can see this completed and available in 2016. Stick to schedule. 2017 is way too far off, and a great opportunity could be lost.
http://www.anandtech.com/show/6129/apple-a4a5-designer-k8-lead-architect-jim-keller-returns-to-amd

 

[Dent1]

AMD's Zen Architecture - Built from the Ground Up.
http://en.wikipedia.org/wiki/AMD_Zen

14nm AMD Zen CPU Will Have DDR4 and Simultaneous Multithreading
http://news.softpedia.com/news/14nm...-and-Simultaneous-Multithreading-471401.shtml

Lets not forget AMD has Jim Keller, most likely the greatest CPU architect in the world. Based on Intel, Apple, Microsoft, Nvidia and IBM.

Zen is being designed to not only compete with Intel's future CPU's, its being designed to overtake them. Never underestimate a company that's been backed into a corner for so long.
After his work on K8, Keller left AMD in 1999. That is when the company started to go down hill.

He's been working on this ZEN project for some time now. I can see this completed and available in 2016. Stick to schedule. 2017 is way too far off, and a great opportunity could be lost.
http://www.anandtech.com/show/6129/apple-a4a5-designer-k8-lead-architect-jim-keller-returns-to-amd

You waited a year to reply to a dead thread. Couldn't you have started a fresh thread?

 

[ThE_MaD_ShOt]

If all comes true I see what my next main rig build is going to be.

 

[xfia]

You waited a year to reply to a dead thread. Couldn't you have started a fresh thread?

less than 6 months.. did we adopt some new calender?

 

[RealNeil]

Zen sounds good and I hope it works out for them. Jim Keller should be able to make the difference.
I don't think that they have to dominate the market to stay viable. It's a huge, multi-layered market and there is a lot of it to go around.

Without them, we would all suffer with higher prices overall.

If all comes true I see what my next main rig build is going to be.

I have two decent Intel Gamers now, and I won a FX-9590 and Board over the holidays. (building this one soon) So I'll have that system too.
When AMD releases all of their latest goodies, I'll be reading reviews, and I will buy if it's warranted.

 

[HumanSmoke]

Lets not forget AMD has Jim Keller, most likely the greatest CPU architect in the world. Based on Intel, Apple, Microsoft, Nvidia and IBM.

ROFL

 

[R-T-B]

ROFL

I don't know enough either way, but if he designed K8 he certainly doesn't suck. And as much as I love to bash apple, their ARM CPUs A5 and such aren't that bad either.

Care to elaborate?

 

[Batou1986]

AMD should be barred from using the words "high performance" since that only applies to their CPU's when intel is not in the comparison.

 

[TRWOV]

I don't think AMD will ever reach IPC parity with Intel but if they manage to match, IDK, Gulftown and keep TPD in check while offering low prices they could do fine.

Their Jaguar cores already outmatch the old A64 IPC (going by Intel Burn Test results) so if they manage to do something similar for their big cores that would be a good start.

 

[Ebo]

I really hope that AMD pulls a rabbit out of the hat.

Not only have they gotten the right people together in the develop team, but have made det design of the chip, from the ground up. That alone should end the poor singletread preformance and make a more effective CPU.

I think we can all agree on that AMD has to be there, not only to make a good chip, but also to keep pushing Intel, så they cant sit with their foot up there ass, which their last 2-3 chips has been.

If AMD deliver a very good chip with Zen, I have no problem building another AMD machine again.

 

[R-T-B]

AMD should be barred from using the words "high performance" since that only applies to their CPU's when intel is not in the comparison.

Someone forgot about the Pentium 4 era.