News Intel Promotes Raja Koduri to Executive VP Following Arc GPU Development

[Admin]

Intel CEO Pat Gelsinger promoted the head of Intel's Accelerated Computing Systems and Graphics, Raja Koduri, to the role of Executive Vice President - boding well for the company's high-performance Arc Alchemist graphics cards.

Intel Promotes Raja Koduri to Executive VP Following Arc GPU Development : Read more

 

[hotaru251]

apple, amd, intel...all thats left for his resume is nvidia.

 

[saltweaver]

Go Raja. I'm hoping Arc succeeds on mid range, enthusiast and pro level.

 

[waltc3]

If you'll check you will see that Intel, like most large companies, has a bunch of Executive Vice Presidents... The timing on this is especially interesting as Intel is yet to ship the discrete versions of ARc and the initial version is, well, not really competitive, AFAIK. Definitely bottom-tier stuff so far.

 

[waltc3]

apple, amd, intel...all thats left for his resume is nvidia.

It's called the "peter principle"...

 

[Giroro]

Isn't this the guy who recycled GCN for like 10 years, being solely responsible for AMD's obsolete GPUs becoming completely uncompetitive with Nvidia (until they finally got rid of him).

Maybe this is a case of him failing upward, because his career so far doesn't suggest he has the slightest idea how to develop a new product, especially not a GPU. He's the kind of uncreative guy who seems more qualified to head up the skylake refresh refresh refresh refresh on 14nm+++++++++. This promotion probably comes with a bigger golden parachute. That might not the best business decision for Intel to make while they are trying to open new markets and struggling to retake the lead in their golden goose server CPUs.

 

[guru7of9]

Raja Koduri's promotion seems a little premature, considering nothing of note has really been released yet from Intel graphics division.
They are celebrating on the assumption its a job well done without any confirmation!
They could well end up with egg on their face in 12 months time!
I guess they are just happy he could improve upon their current igp's, as their last foray into discreet graphics was a dismal failure 15 odd years ago.
In essence they set the bar pretty low !
But we shall see !

 

[watzupken]

Raja Koduri's promotion seems a little premature, considering nothing of note has really been released yet from Intel graphics division.
They are celebrating on the assumption its a job well done without any confirmation!
They could well end up with egg on their face in 12 months time!
I guess they are just happy he could improve upon their current igp's, as their last foray into discreet graphics was a dismal failure 15 odd years ago.
In essence they set the bar pretty low !
But we shall see !

I was thinking the same. It is like, “Congrat! You are promoted for a product that is yet to launch”. I won’t count DG1 as a successful launch. It was “released” quietly, and seems to have vanished under the radar.

 

[watzupken]

Isn't this the guy who recycled GCN for like 10 years, being solely responsible for AMD's obsolete GPUs becoming completely uncompetitive with Nvidia (until they finally got rid of him).

Maybe this is a case of him failing upward, because his career so far doesn't suggest he has the slightest idea how to develop a new product, especially not a GPU. He's the kind of uncreative guy who seems more qualified to head up the skylake refresh refresh refresh refresh on 14nm+++++++++. This promotion probably comes with a bigger golden parachute. That might not the best business decision for Intel to make while they are trying to open new markets and struggling to retake the lead in their golden goose server CPUs.

Objectively, I feel he don’t have a lot of budget then to actively make changes to improve GPUs. It wasn’t so long ago that AMD almost went bankrupt. So perhaps he may fare better over at Intel with fat budgets. The problem here is that he is promoted for his effort for creating the ARC dGPUs, which ironically has been delayed again and again, and not yet available for sale.

 

[escksu]

If you'll check you will see that Intel, like most large companies, has a bunch of Executive Vice Presidents... The timing on this is especially interesting as Intel is yet to ship the discrete versions of ARc and the initial version is, well, not really competitive, AFAIK. Definitely bottom-tier stuff so far.

You need to realise that the main aim for Intel's GPU isn't really about gaming GPUs. The real aim is HPC-GPU, i.e. Ponte Vecchio. HPC-GPUs are highly sought after esp. in supercomputers. This is where Intel is heading.

 

[escksu]

Objectively, I feel he don’t have a lot of budget then to actively make changes to improve GPUs. It wasn’t so long ago that AMD almost went bankrupt. So perhaps he may fare better over at Intel with fat budgets. The problem here is that he is promoted for his effort for creating the ARC dGPUs, which ironically has been delayed again and again, and not yet available for sale.

ARC will be delayed for a long long time... IT won't be available till Intel gets Ponte Vecchio out. Thats the real purpose why Intel even wants to revive its GPU dept.

 

[escksu]

I was thinking the same. It is like, “Congrat! You are promoted for a product that is yet to launch”. I won’t count DG1 as a successful launch. It was “released” quietly, and seems to have vanished under the radar.

DG1 is nothing more than a proof of concept.... Intel never intend to compete with Nvidia and AMD on the gaming front.... They are after the HPC segment. This is where the real money is. With Ponte Vecchio, Intel could now supply both CPU and GPUs are HPC sector, esp. supercomputers and other high end servers dedicated to things like AI, deep learning etc...

 

[watzupken]

DG1 is nothing more than a proof of concept.... Intel never intend to compete with Nvidia and AMD on the gaming front.... They are after the HPC segment. This is where the real money is. With Ponte Vecchio, Intel could now supply both CPU and GPUs are HPC sector, esp. supercomputers and other high end servers dedicated to things like AI, deep learning etc...

I don’t disagree. However my point was there is no reason for the promotion based on a paper launch of the ARC GPUs. Even Ponte Vecchio is nothing more than a bunch of pictures with Raja holding the chip in his hands at this point in time.

 

[TerryLaze]

DG1 is nothing more than a proof of concept.... Intel never intend to compete with Nvidia and AMD on the gaming front.... They are after the HPC segment. This is where the real money is. With Ponte Vecchio, Intel could now supply both CPU and GPUs are HPC sector, esp. supercomputers and other high end servers dedicated to things like AI, deep learning etc...

Last year intel made $40.5 billion from client and $25.8 billion from data center...and it's not last year only either this has been going on for many years, client is almost double the business from data centers for intel.
Intel releasing DG1 first was 100% intentional and their biggest priority, the more OEMs they can convince to use DG1 instead of crap gt710 cards or whatever other crap OEMs can get their hands on, the more money intel will be making from the systems that are already being sold anyway.

Investor Relations

Intel (Nasdaq: INTC) is an industry leader, creating world-changing technology that enables global progress and enriches lives. Inspired by Moore’s Law, we continuously work to advance the design and manufacturing of semiconductors to help…

www.intc.com

 

[jp7189]

Isn't this the guy who recycled GCN for like 10 years, being solely responsible for AMD's obsolete GPUs becoming completely uncompetitive with Nvidia (until they finally got rid of him).

Maybe this is a case of him failing upward, because his career so far doesn't suggest he has the slightest idea how to develop a new product, especially not a GPU. He's the kind of uncreative guy who seems more qualified to head up the skylake refresh refresh refresh refresh on 14nm+++++++++. This promotion probably comes with a bigger golden parachute. That might not the best business decision for Intel to make while they are trying to open new markets and struggling to retake the lead in their golden goose server CPUs.

Agree. Notice how AMD became made a huge leap in GPU after Koduri was out of the way. I think it's pretty clear Korudi is behind Intel's pivot from the upcoming low-profit GPU flop to an extremely profitable (environment destroying) bitcoin miner.

 

[digitalgriffin]

I like Raja, but he's always been a day late and dollar short after the 280X. The pattern is repeating for Intel. That in my opinion doesn't deserve promotion until he comes up with something competing on time.

 

[Deleted member 14196]

well, it's like the army, when you screw up they promote you to get rid of you.. it was referred to above as failing upwards. it's not what you know, it's who you know--and sometimes who you blow

i am not impressed by any of them to be honest.

 

[jacob249358]

maybe he made some decent gpus

 

[saltweaver]

DG1 is nothing more than a proof of concept.... Intel never intend to compete with Nvidia and AMD on the gaming front.... They are after the HPC segment. This is where the real money is. With Ponte Vecchio, Intel could now supply both CPU and GPUs are HPC sector, esp. supercomputers and other high end servers dedicated to things like AI, deep learning etc...

Does Intel have solution for GPUs like AMD Radeon PRO or Nvidia Quadro ?

 

[Alpha_Lyrae]

Isn't this the guy who recycled GCN for like 10 years, being solely responsible for AMD's obsolete GPUs becoming completely uncompetitive with Nvidia (until they finally got rid of him).

Maybe this is a case of him failing upward, because his career so far doesn't suggest he has the slightest idea how to develop a new product, especially not a GPU. He's the kind of uncreative guy who seems more qualified to head up the skylake refresh refresh refresh refresh on 14nm+++++++++. This promotion probably comes with a bigger golden parachute. That might not the best business decision for Intel to make while they are trying to open new markets and struggling to retake the lead in their golden goose server CPUs.

It's easy to bash Raja at AMD (especially after Vega), but he was the primary architect (among a team of others) of Navi/RDNA, which ironically enough, was AMD's return to competitiveness. Architects, inventors, and engineers are often working on products 2-4 years ahead of retail (from conceptualization to working silicon samples).

AMD's lean years were rough. Consoles and dwindling PC market share barely kept them afloat. Raja had two stints at ATI/AMD from 2001-2009, then 2013-2017. VLIW2 RDNA ("Super SIMD") was patented in 11/2016. Vega's SoC design with Infinity Fabric was apparently difficult, but that also laid the groundwork for RDNA's on-chip network. GCN1 hit retail in 2012. Raja returned around the time of 2013 GCN2 (Hawaii), then was around for: 2014
/2015 GCN3 (Tonga/Fiji), 2015's mess of GCN1-2 rehashes, 2016 GCN4 (Polaris), 2017 GCN5 (Vega). So, no, not 10 years.

Vega's ultimate failure was its power consumption and primitive shaders that were touted, but never delivered. Worse yet, the whitepaper listed 17x geometry culling speedup was likely a very specific case in SPECperf or 3D rendering application (not gaming). RDNA's geometry culling improved by 2x over GCN via primitive shaders, not unobtainium 17x. So, yeah, overpromising and failing to deliver features didn't help his case.

AMD also had to get rid of Raja because he kept wanting RTG sold to Intel, which was never going to happen. There were also rumors that he and Dr. Su butted heads more often than not, though that can't really be confirmed.