Apple management was far better than rory reads, lisa sus, hector ruizes, papermasters and bruce claflins that sit on AMD BoD.
Apple management was far better than rory reads, lisa sus, hector ruizes, papermasters and bruce claflins that sit on AMD BoD.
I am a huge AMD fan, 3 out of my 5 builds have been AMD. But I cant justify getting such low performing CPUs.
Not a flame bait thread, I really want to proudly buy a AMD CPU and tell the world to buy AMD and that its BETTER than intel like during the Athlon64 days. I honestly cant recommend an AMD unless they are on a really tight budget.
Yes, even when outperforming Intel s CPUs they re still low performing CPUs...
http://www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html
I guess that for some people numbers do not matters when checking a CPU perf, surely that the brand is a more accurate benchmark after all, and of course hearsay, the best bench to this day assuming one get the infos in the "good" forums...
myocardia
Diamond Member
- Jun 21, 2003
- 9,291
- 30
- 91
??? Yet the 88 watt 4790k utterly destroys the 220 watt FX 9590 on the page you linked above...
My point is price/perf and in this respect the 4790K destroy the 9590 price wise but certainly not in perfs, there s 10% difference, i guess that it s enough for some to be pointed as huge..
Seriously.??.
What is the price difference.?. 10%..?.
The graph point the FX8350 as the best value of the bunch, here it is currently 180, a 4670K is not as fast and is 246, for this price you have a 8350 and a decent MB...and more perfs!!...
A FX8320 is 150, at this level a i5 is a joke perf/Watt wise since a 3.1 is 186 for instance.
Phynaz
Lifer
- Mar 13, 2006
- 10,140
- 819
- 126
Don't bother man. You'll get better results hitting yourself in the head with a hammer.
frozentundra123456
Lifer
- Aug 11, 2008
- 10,451
- 642
- 126
The same thread crappers unrelentlessly crapping AMD related threads, not the slightest technical point in their posts, that s all ad hominems from the usual trolls...
Phynaz
Lifer
- Mar 13, 2006
- 10,140
- 819
- 126
There's a lot of fanboyism in this thread. Yes it's true that Intel offers better single core performance. AMD isn't as far behind as people like to believe. My specs are as follows:
Processor: AMD FX-6300 OC'd to 4.3 GHz
RAM: Adata XPG 8GB GDDR3 1600 MHz
Mobo: Asus M5A78L-M LX Plus
GPU: Asus GTX 770 OC 2GB
PSU: EVGA Supernova 650W 80+ Gold
Storage: Adata 256GB SSD
Monitor: BenQ 24" 1080P 60Hz w/ 1ms response time
My specs are decent, my build cost about $1100 USD. I ran benchmarks last night, unless otherwise noted, graphics are maxed out at 1080P with Vsync enabled. In Skyrim I show cores 0 and 1 at ~45-60% load and the rest hover at around ~20-25% load. In GW2 I see similar figures, cores 3 and 4 are around 50-70% load while the rest fluctuate around ~20%. In Call of Duty: Advanced Warfare, the story is a bit different. The game is well threaded, all cores revolve around ~60-70% load. My GPU on the other hand sits at a constant 98-99% load in COD:AW which tells me it's time for an upgrade even though I've got the game on high settings w/ Vsync and it does run smoothly for the most part. I'm sure I could get BETTER performance by switching to an i5 4690K however my frames per second aren't going to improve that much more. We can theory-craft and benchmark all day long, what it comes down to is real life performance. I'm able to pull a constant 60FPS with all the games I've mentioned either maxed out or almost maxed out. If you're gonna drop $2,000+ on a 4K gaming rig sure go Intel, it's the sensible thing to do. If you're on a budget and want good performance at a fair price, AMD won't let you down. I'm sure if I upgrade to an FX-8350 and OC that kitten my eyes (not benchmarks but my eyes) won't be able to tell the difference from an Intel Core i5. Just my $0.02
Processor: AMD FX-6300 OC'd to 4.3 GHz
RAM: Adata XPG 8GB GDDR3 1600 MHz
Mobo: Asus M5A78L-M LX Plus
GPU: Asus GTX 770 OC 2GB
PSU: EVGA Supernova 650W 80+ Gold
Storage: Adata 256GB SSD
Monitor: BenQ 24" 1080P 60Hz w/ 1ms response time
My specs are decent, my build cost about $1100 USD. I ran benchmarks last night, unless otherwise noted, graphics are maxed out at 1080P with Vsync enabled. In Skyrim I show cores 0 and 1 at ~45-60% load and the rest hover at around ~20-25% load. In GW2 I see similar figures, cores 3 and 4 are around 50-70% load while the rest fluctuate around ~20%. In Call of Duty: Advanced Warfare, the story is a bit different. The game is well threaded, all cores revolve around ~60-70% load. My GPU on the other hand sits at a constant 98-99% load in COD:AW which tells me it's time for an upgrade even though I've got the game on high settings w/ Vsync and it does run smoothly for the most part. I'm sure I could get BETTER performance by switching to an i5 4690K however my frames per second aren't going to improve that much more. We can theory-craft and benchmark all day long, what it comes down to is real life performance. I'm able to pull a constant 60FPS with all the games I've mentioned either maxed out or almost maxed out. If you're gonna drop $2,000+ on a 4K gaming rig sure go Intel, it's the sensible thing to do. If you're on a budget and want good performance at a fair price, AMD won't let you down. I'm sure if I upgrade to an FX-8350 and OC that kitten my eyes (not benchmarks but my eyes) won't be able to tell the difference from an Intel Core i5. Just my $0.02
Last edited:
podspi
Golden Member
- Jan 11, 2011
- 1,982
- 102
- 106
This can trivially be said about almost anything that isn't a commodity. That statement is loaded with a bunch of assumptions that you do not support. Not one person on this board has actually even attempted, even touched on the fact that saying there can only be a monopoly in semiconductors implies that the minimum efficient scale in semiconductors is huge compared to the size of the overall market. This might be true, but since nobody has mentioned that, let alone tried to show it, suggests to me that at the very least nobody on these boards has any proof of it.
Which means we should stop repeating it, and we should start demanding actual proof for what is a ridiculously strong assertion that isn't backed by reality (there are, currently, multiple semiconductor companies developing high performance nodes and ICs that make a profit) when others repeat it.
It might surprise a lot of people on these forums that there are industries that spend even MORE on R&D selling differentiated but similar products and still have competition. Fixed costs do not imply a monopoly needs to be the result - especially in a market like semiconductors where products are differentiated
Edit:
Added links.
And yeah that Forbes article (2nd link) isn't the best - I am just looking at the R&D #s.
Last edited:
myocardia
Diamond Member
- Jun 21, 2003
- 9,291
- 30
- 91
You never mentioned price in the post that I quoted. It makes more sense, now that you have clarified.
DrMrLordX
Lifer
- Apr 27, 2000
- 23,222
- 13,300
- 136
Oh stop hectoring the man!
ba dum, tsh.
There is more to the story than simply "AMD can't compete". AMD has, since ~2011, produced CPUs that perform fairly well compared to Intel offerings in edge cases that either aren't significant enough to give AMD a market advantage or are too difficult for anyone to produce outside of a very small niche.
Bulldozer and Piledriver are actually pretty good in integer performance (some of the time), but that didn't make a big enough difference for them to catch on in the server room, so the floor sort of dropped out from under the entire company budget-wise. There was also the whole FMA3/FMA4 thing . . .
Then, after that, AMD made a push for HSA with Kaveri (which showed up late), and nobody (including AMD!) supported it with software to take advantage of its capabilities at launch. So here's an AMD chip that brand new sold for ~$180 that could challenge many Intel chips (even a few Haswells) with a bunch of dark silicon because nobody could use it. Even today, HSA doesn't work under anything but Linux, and even then, high-level language support for HSA features is a work in progress.
It's not like AMD hasn't put forth CPU designs that have a lot of performance potential. They have. If you take full advantage of xOP, FMA4, and HSA, Kaveri can stack up quite well against Intel offerings, even if FMA3 and AVX2 are also supported (and what the heck, even OpenCL for HD4xxx GPGPU where possible). Most developers aren't going to bother with that, and absolutely nobody is supporting HSA right now that I can tell.
What you usually get is some code with SSE2/SSE3 support and that's it, and when it's code like that (and/or legacy code), Intel wins. When it's sloppy code that causes cache thrashing and other madness, Intel grabs a bigger lead (and HT often helps with those problems in multithreaded situations). AMD's latest offerings are simply too dependent on support from the developer and the toolchain.
You never mentioned price in the post that I quoted. It makes more sense, now that you have clarified.
Sorry for being unclear....
Indeed, with a dwindling , prices have become quite an issue here in Europe with about all gear being 20% more expensive than just 6-8 months ago.
A FX8350 is 180 while it was 149, Intel s 4670K made the same jump %age wise and is now in the 240+ zone, that s not a good evolution for customers as the same PC was significantly cheaper last year, and in these conditions there s surely a lot of people not upgrading or not buying new PCs because of prices perceived as excessive.
If you're that concerned about price then you have to factor power consumption as well. Intel is exceeding AMD's performance while using substantially less power (50% or more). Electricity isn't free, or at least isn't on this side of the pond. If you're planning on keeping this system a while, that adds up.
CHADBOGA
Platinum Member
- Mar 31, 2009
- 2,135
- 833
- 136
Server & Desktop marketshare shows how wrong you are, yet from this position of confusion, you seek to lecture others.
No one in their right mind wants a 220w space heater, unless they live in freezing conditions.
The lack of single core performance that the FX series offers, means that one will run into problems playing future games, long before someone owning a 4 core or greater i5 or i7, will.
It is for the above reasons(which you continue to fail to comprehend), why AMD's sales of the FX series are so uninspiring.
To add to that, does anyone seriously think that when AMD were designing Bulldozer they were doing it with the goal of being a cheaper alternative to Sandy/Ivy Bridge, with far higher power usage and performance that was only really competitive with its Intel rivals in certain heavily multithreaded scenarios? Or likewise, that their intention with the original Phenom was to be a slow, dirt-cheap alternative to Penryn quad-cores, and a drop-in upgrade for Athlon 64 X2s that caused you to lose a lot of single-thread performance?
AMD's problem over the last few years has been designing chips that were far too ambitious in their scope, and then proving unable to execute them. Insisting on a 4-core die across the entire Phenom range was no doubt intended to show their engineering superiority to Intel, but it actually ended up showing the opposite. As for Bulldozer, it was pretty much the equivalent of taking out three mortgages on your house, then going to Vegas and betting all the money on red. And it came up black.
AMD's problem over the last few years has been designing chips that were far too ambitious in their scope, and then proving unable to execute them. Insisting on a 4-core die across the entire Phenom range was no doubt intended to show their engineering superiority to Intel, but it actually ended up showing the opposite. As for Bulldozer, it was pretty much the equivalent of taking out three mortgages on your house, then going to Vegas and betting all the money on red. And it came up black.
ShintaiDK
Lifer
- Apr 22, 2012
- 20,378
- 146
- 106
Yep. Its a company out of touch with reality and its capabilities. Servers and arrogance is the main culpit. The Bulldozer uarch for example is an uarch specificly designed for servers, yet they have no marketshare there to speak of. It was also a speedracer design to make it even worse. Like no lesson was learned from Intels Netburst case. In terms of the native quadcore, funny that AMD later adopted the MCM approach they raged in marketing against. Just in the server space tho.
And that leads us back to the future. Again they are going for the server CPU and way overextending their capabilities.
Last edited:
Bulldozer was a clear message to the world that they have given up the race against Intel and just went all bonanza with power and cores.
3-4 years in development and that was what they came up with?
3-4 years in development and that was what they came up with?
AMD is hoping multi threading will start to be the norm, and Intel guessed single thread was the answer.
Intel guessed it right.
You can`t just scale up the chips to one gigantic die, and hope that will do well in the public eyes against a smaller and much more efficient chip
Even if the interest is there, the only way AMD would be able to sell these chips is through dumping the price and reducing their margins, and effectively cutting off a huge part of their R&D because that cost money. Which turns in to a vicious cycle
You can`t just scale up the chips to one gigantic die, and hope that will do well in the public eyes against a smaller and much more efficient chip
Even if the interest is there, the only way AMD would be able to sell these chips is through dumping the price and reducing their margins, and effectively cutting off a huge part of their R&D because that cost money. Which turns in to a vicious cycle
Designing for servers first was an approach that had worked well for them with the Athlon 64/Opteron, and previously also worked for Intel with the Pentium Pro (once they had tweaked the design with the Pentium II). The problem was that the needs of the server and client side started to diverge as the 2000s went along, as evidenced by the fact that K10 walked all over the LGA771 Xeons once they'd gotten the clockspeed up, yet its native quad-core design offered little advantage over the Core 2 Quad, even leaving out the fact that Intel had much higher clockspeeds and had moved onto Penryn. In fact, even Nehalem was obviously a server-oriented design, where it completely obliterated AMD's performance advantage, but scaled down poorly and resulted in the hackjob solution that was Clarkdale for the lower-end.
What I think really hurt AMD, possibly even as much as Core 2 did, was Sandy Bridge. In one go, Intel came up with a single basic design that worked just as well for laptops, desktops and servers alike, while offering single-thread performance leagues ahead of what AMD could offer. Even if Bulldozer had delivered at the higher-end, Llano and the later FX-based APUs would still likely have been outperformed at a lower power usage by Sandy Bridge and its successors.
el etro
Golden Member
- Jul 21, 2013
- 1,584
- 14
- 81
Yeah. And CMT alone failed, failed on theory and much more failed in the AMD implementation.
Why? What personal connection do you have to AMD that makes you want to do this? Or disdain of intel?
Can't agree more with this statement.
Although I also think the designs were far ahead/not in line with consumers desires. Massive multicore designs without the actual use cases for those designs was a massive mistake.
Last edited:
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 24K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter