Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [DT] AMD updates product roadmap for 2014 and 2015
New Posts  All Forums:Forum Nav:

[DT] AMD updates product roadmap for 2014 and 2015 - Page 8

post #71 of 91
Quote:
Originally Posted by fateswarm View Post

SCollins if you spent more time knowing about those things instead of trying to fish an ad hominem attack, you'd know that concurrency must include context switches.

I think your talking out of your rear facing sphincter mostly, and my point was that context switchs cost nothing in terms of latency.

I also pointed you to massively multithreaded APIS that focus on low latency, in fact multithreading in most sitautions is great at reducing latencys if applied with some sense.
post #72 of 91
Quote:
Originally Posted by egkv View Post

Well, you did make very wrong statement and claimed your "two years of hacking id3tech" to justify it. That looked incredibly silly from my point of view - as an actual software engineer who spent last 10 years in heavily multithreaded world.
Please define "non-interactive" as opposed to "interactive".

I think he might be reffering to terminal no gui applications and Gui non terminal applications.

but all aplications are interactive for the most part if they workd with changing variables. The only style of non interactive applications would be something like a calculator that is preprogrammed to run one routine for a solutions, once.

I think either his english sucks or he is just trolling people. I think we should stop feeding the troll
post #73 of 91
Quote:
Originally Posted by fateswarm View Post

SCollins if you spent more time knowing about those things instead of trying to fish an ad hominem attack, you'd know that concurrency must include context switches.
Wait, what? As in - CPU-level context switching? Could you clarify, please, how is this relevant to multicore hardware? I'd also like to know how OS events and hardware interrupts do not cause context switching in single-threaded applications.
post #74 of 91
Imo the whole threaded vs non threaded latency argument is nonsense. Utilizing one or more threads creates no latency based on how you pipe your resources. In most cases applications use shared memory. Access times to memory are so instantaneous that you cant really call it a delay. The only real "latency" is how long it takes each thread to complete its loop before it's back to accessing the memory. Any possible "latency" created (there is none for me) is more than made up ten fold of how much faster multiple threads can complete a large task in comparison to a single thread.
The Beast
(13 items)
 
Sero 7 Pro
(8 items)
 
 
CPUMotherboardGraphicsRAM
A10-6800k ASRock FM2A85X Extreme6 ASUS HD 5870 8GB AMD Radeon™ RP1866 
Hard DriveOptical DriveCoolingOS
60GB SSDNow V300 LG 24X FX Liquid Loop Windows 8 x64 
MonitorKeyboardPowerCase
HP 2009m Logitech K120 Fractal Tesla R2 650W Fractal Arc Midi R2 
Mouse
HP Modguo 
CPUGraphicsRAMHard Drive
Nvidia Tegra3 ULP GeForce 1 GB 8 GB 
OSMonitorPowerAudio
Android 4.2 (Jelly Bean) 7" IPS @ 1280×800 4000 mAh Stereo Speakers 
  hide details  
Reply
The Beast
(13 items)
 
Sero 7 Pro
(8 items)
 
 
CPUMotherboardGraphicsRAM
A10-6800k ASRock FM2A85X Extreme6 ASUS HD 5870 8GB AMD Radeon™ RP1866 
Hard DriveOptical DriveCoolingOS
60GB SSDNow V300 LG 24X FX Liquid Loop Windows 8 x64 
MonitorKeyboardPowerCase
HP 2009m Logitech K120 Fractal Tesla R2 650W Fractal Arc Midi R2 
Mouse
HP Modguo 
CPUGraphicsRAMHard Drive
Nvidia Tegra3 ULP GeForce 1 GB 8 GB 
OSMonitorPowerAudio
Android 4.2 (Jelly Bean) 7" IPS @ 1280×800 4000 mAh Stereo Speakers 
  hide details  
Reply
post #75 of 91
Quote:
Originally Posted by SCollins View Post

I think he might be reffering to terminal no gui applications and Gui non terminal applications.
But every reasonably programmed GUI application is at least two-threaded - one being UI thread that controls interface repaint routines and while another does actual work. There are some terrible purely single-threaded programs out there, you know it when it completely freezes (can't drag / resize window, mouse leaves traces on it etc) after you have clicked a button and until it finishes some internal calculations. There were much more of them back in Windows 95 times - usually written by hobbyists with Delphi or similar "RAD" tool.[/quote]
post #76 of 91
Fateswarm knows everything, don't try to argue a point.

Nvidia is the best, gtx 680 not the same as a 770.

and whatever drivel he is going on about ^ there, he is right.

rolleyes.gif
Gaming Rig
(15 items)
 
Sons Rig
(11 items)
 
File Server
(13 items)
 
CPUMotherboardGraphicsGraphics
AMD Ryzen 1700x ASUS Prime X370 Pro Gigabyte GeForce GTX 980 Gigabyte GeForce GTX 980 
RAMHard DriveCoolingOS
16gb's Team Select 3200Mhz RGB  Crucial M500 SSD 240gb Corsair H55 Windows 7 Ultimate 
MonitorMonitorPowerCase
ASUS PB287Q 4K Monitor HTC Vive Coolermaster Silent Pro M 1000W Rosewill Blackhawk Ultra 
MouseMouse PadAudio
Razer Deathadder 2013 Steelseries Qck Mass Super Thick Cloth Mouse Pad Genius SW-G2.1 1250 4PC Gaming Speakers 
CPUMotherboardGraphicsGraphics
Core i7 3770k Asus P8Z68 Deluxe Galaxy Geforce GTX 780 Galaxy Geforce GTX 780 
RAMHard DriveCoolingOS
16 Gb's G.Skill DDR3 1866 Crucial M500 SSD 240Gb Coolit ECO C240 Windows 7 Ultimate 
MonitorPowerCase
Benq G2420HD Coolermaster Silent Pro 1000 Watt Deepcool Kendomen 
CPUMotherboardRAMHard Drive
Core i3 2120 Supermicro X9SCM-F 4 Gb Kingston 1600mhz DDR3 ECC 12 Segate 2tb drives (RAID 6) 
Hard DriveHard DriveOSPower
ADATA SP600 SSD 24 Toshiba DT01ACA200 drives (2 RAID 6's) Windows 7 Ultimate 64 bit. OCZ ZT750 Supermicro PWS-665-PQ 
CaseOtherOtherOther
Norco 4020 and Norco 4224 LSI 9260-4i Raid Card Intel RES2SV240 20 port Expander. HP SAS 24 por... Voltaire 410Ex Hca Infiniband HBA 
  hide details  
Reply
Gaming Rig
(15 items)
 
Sons Rig
(11 items)
 
File Server
(13 items)
 
CPUMotherboardGraphicsGraphics
AMD Ryzen 1700x ASUS Prime X370 Pro Gigabyte GeForce GTX 980 Gigabyte GeForce GTX 980 
RAMHard DriveCoolingOS
16gb's Team Select 3200Mhz RGB  Crucial M500 SSD 240gb Corsair H55 Windows 7 Ultimate 
MonitorMonitorPowerCase
ASUS PB287Q 4K Monitor HTC Vive Coolermaster Silent Pro M 1000W Rosewill Blackhawk Ultra 
MouseMouse PadAudio
Razer Deathadder 2013 Steelseries Qck Mass Super Thick Cloth Mouse Pad Genius SW-G2.1 1250 4PC Gaming Speakers 
CPUMotherboardGraphicsGraphics
Core i7 3770k Asus P8Z68 Deluxe Galaxy Geforce GTX 780 Galaxy Geforce GTX 780 
RAMHard DriveCoolingOS
16 Gb's G.Skill DDR3 1866 Crucial M500 SSD 240Gb Coolit ECO C240 Windows 7 Ultimate 
MonitorPowerCase
Benq G2420HD Coolermaster Silent Pro 1000 Watt Deepcool Kendomen 
CPUMotherboardRAMHard Drive
Core i3 2120 Supermicro X9SCM-F 4 Gb Kingston 1600mhz DDR3 ECC 12 Segate 2tb drives (RAID 6) 
Hard DriveHard DriveOSPower
ADATA SP600 SSD 24 Toshiba DT01ACA200 drives (2 RAID 6's) Windows 7 Ultimate 64 bit. OCZ ZT750 Supermicro PWS-665-PQ 
CaseOtherOtherOther
Norco 4020 and Norco 4224 LSI 9260-4i Raid Card Intel RES2SV240 20 port Expander. HP SAS 24 por... Voltaire 410Ex Hca Infiniband HBA 
  hide details  
Reply
post #77 of 91
Quote:
Originally Posted by Opcode View Post

Imo the whole threaded vs non threaded latency argument is nonsense. Utilizing one or more threads creates no latency based on how you pipe your resources. In most cases applications use shared memory. Access times to memory are so instantaneous that you cant really call it a delay. The only real "latency" is how long it takes each thread to complete its loop before it's back to accessing the memory. Any possible "latency" created (there is none for me) is more than made up ten fold of how much faster multiple threads can complete a large task in comparison to a single thread.

I kind of understand what my opponent actually had in mind while talking about multithreading latencies - but is has nothing to do with "context switching" whatsoever.
I'm just waiting patiently until he properly defines a problem and provides a foundation for further discussion. Right now he is still fuming about his experience of idtech engine hacking getting heavily discounted, we should let him some time to get over it.
post #78 of 91
I dont understand why AMD is abandoning pure CPUs and going all APU... I mean APUs are great for low-end gaming systems or just general purpose PCs. I have one in the office PC I built this summer, and it serves its purpose well at an extremely good price. But come on, we still need enthusiast CPUs and AMD is apparently just throwing in the white flag and handing Intel a total monopoly on the enthusiast CPU market.

It doesn't make sense to me why AMD, a producer of a very wide range of discrete GPUs, would abandon GPU-less CPU dies. If I already have a discrete GPU or plan to build the rest of my system around one, when it comes to looking for CPUs, I'm more inclined to seek out a GPU-less, pure CPU chip. Why? Because if a die has an IGP, the price of the transistors, silicon, and the research+development of the IGP is factored into the cost of the chip. Since I will never even once use the IGP portion of the die, that's a waste of money and silicon die space for me. With the FX CPUs, while they lacked badly in single-thread performance, at least I knew I wasn't paying extra for parts of the chip I'd never use, and I was getting something in return for every dollar spent on the CPU.

When I bought my i5 3570K, I knew that the insignificant Intel HD 3000 IGP didn't really add much cost to the CPU, but I still wished they offered a GPU-less version for a little lower cost. Still, it's not so bad considering the chip is predominantly a CPU with a little added bonus instead of a half-and-half chip. With APUs, AMD tries to scale the power of the GPU with the CPU so they equally matched up and one doesn't bottleneck the other. This means if they come out with an 8-core APU with great single-thread performance, it must be paired with an equally-powered GPU, which by itself will also cost quite a bit.

Buyers of a top-line APU will lose interest in discrete graphics, effectively shrinking the discrete GPU market, in which AMD is obviously heavily involved in, so they are kind of shooting themselves in the foot here, especially considering the fact that it's highly unlikely the CPU-portion of their high-end APUs will ever catch up to Intel's i7/i5 (or whatever they're called in 2015) CPUs, and then AMD loses on both fronts. It just really makes no sense to me for AMD to go all-APU from this point on... It also makes me scared of what Intel is going to do with its enthusiast-CPU monopoly...
post #79 of 91
Quote:
Originally Posted by iamhollywood5 View Post

I dont understand why AMD is abandoning pure CPUs and going all APU... I mean APUs are great for low-end gaming systems or just general purpose PCs. I have one in the office PC I built this summer, and it serves its purpose well at an extremely good price. But come on, we still need enthusiast CPUs and AMD is apparently just throwing in the white flag and handing Intel a total monopoly on the enthusiast CPU market.

It doesn't make sense to me why AMD, a producer of a very wide range of discrete GPUs, would abandon GPU-less CPU dies. If I already have a discrete GPU or plan to build the rest of my system around one, when it comes to looking for CPUs, I'm more inclined to seek out a GPU-less, pure CPU chip. Why? Because if a die has an IGP, the price of the transistors, silicon, and the research+development of the IGP is factored into the cost of the chip. Since I will never even once use the IGP portion of the die, that's a waste of money and silicon die space for me. With the FX CPUs, while they lacked badly in single-thread performance, at least I knew I wasn't paying extra for parts of the chip I'd never use, and I was getting something in return for every dollar spent on the CPU.

When I bought my i5 3570K, I knew that the insignificant Intel HD 3000 IGP didn't really add much cost to the CPU, I still wished they offered a GPU-less version for a little lower cost. Still, it's not so bad considering the chip is predominantly a CPU with a little added bonus instead of a half-and-half chip. With APUs, AMD tries to scale the power of the GPU with the CPU so they equally matched up and one doesn't bottleneck the other. This means if they come out with an 8-core APU with great single-thread performance, it must be paired with an equally-powered GPU, which by itself will also cost quite a bit.

Buyers of a top-line APU will lose interest in discrete graphics, effectively shrinking the discrete GPU market, in which AMD is obviously heavily involved in, so they are kind of shooting themselves in the foot here, especially considering the fact that it's highly unlikely the CPU-portion of their high-end APUs will ever catch up to Intel's i7/i5 (or whatever they're called in 2015) CPUs, and then AMD loses on both fronts. It just really makes no sense to me for AMD to go all-APU from this point on... It also makes me scared of what Intel is going to do with its enthusiast-CPU monopoly...

From the looks of it and all the high end boards coming out for it, they won't be low end gaming for very long
post #80 of 91
Quote:
Originally Posted by Alatar View Post

Source

This article has a lot of inconsistencies, it seems that a lot of facts is mingled here probably unintentionally because they do not know all details. rolleyes.gif

Quote:
Originally Posted by sdlvx View Post

Yesterday I saw a post on Tom's forum where a member asked someone with AMD about Steamroller FX, and the response was pretty much "I can't say anything now but keep the faith."

I have a feeling AMD is going to do like Intel does where they release mobile oriented parts and then a year later or so they release the big ones, so they can focus on high growth areas. I am kind of expecting AM3+ and whatever comes after that (in my dreams a unified server/desktop socket) to end up like LGA2011, where the architecture is behind by a year or so.

25w quad with GCN GPU is pretty impressive though. But it does mean the delays are official. Hopefully I can snag one of these before I leave for my travels in April, I hate carrying around my giant core 2 duo abomination.

But today we have Quad Core Kabini APU A6-5200/GCN GPU/25W TDP, and for those who do not know 25W TDP this is not a power consumption number this is TDP value.I do not understand why only mentioned, a 25W power consumption why mention this in such way it seems that they are confusing TDP and power consumption numbers no doubt.

http://www.cpu-world.com/CPUs/Jaguar/AMD-A6-Series%20A6-5200.html


Current Kabini APU A4-5000/GCN GPU/15W TDP is also a Quad Core, the new Kabini models for 2014 mentioned A4-5350 and A4-5150 they will have higher CPU frequency and for the first time Turbo Core on Jaguar CPU. smile.gif

http://www.cpu-world.com/CPUs/Jaguar/AMD-A4-Series%20A4-5000.html
AMD Steam Engine
(13 items)
 
  
CPUMotherboardGraphicsRAM
AMD Phenom II X6 1055T Asrock A770DE+ Powercolor Radeon HD 3650 Kingmax Mars ddr2 800mhz, 3GB 
Hard DriveOptical DriveCoolingOS
Samsung HD502JH 500GB LG GSA H-30N CPU Cooler Spire Coolgate 2011 Windows 7 Ultimate 
MonitorKeyboardPowerCase
Benq GW2760HS Logitech Media Keyboard 600 Seasonic S12II 430W Gigabyte GZ-X2 Black 
Mouse
Trust Laser 
  hide details  
Reply
AMD Steam Engine
(13 items)
 
  
CPUMotherboardGraphicsRAM
AMD Phenom II X6 1055T Asrock A770DE+ Powercolor Radeon HD 3650 Kingmax Mars ddr2 800mhz, 3GB 
Hard DriveOptical DriveCoolingOS
Samsung HD502JH 500GB LG GSA H-30N CPU Cooler Spire Coolgate 2011 Windows 7 Ultimate 
MonitorKeyboardPowerCase
Benq GW2760HS Logitech Media Keyboard 600 Seasonic S12II 430W Gigabyte GZ-X2 Black 
Mouse
Trust Laser 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Rumors and Unconfirmed Articles
Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [DT] AMD updates product roadmap for 2014 and 2015