AMD's new Radeon RX 3080 XT: RTX 2070 performance for $330?

[Neftex]

sure boys, i can only imagine how would the rest of the lineup sell with worse price/perf leaving the "halo" product the same. the rest of the lineup would get shit on

[pas008]

31 minutes ago, Neftex said:

sure boys, i can only imagine how would the rest of the lineup sell with worse price/perf leaving the "halo" product the same. the rest of the lineup would get shit on

nobody is arguing price to performance having a factor

we been stating many times over best flagship/halo products also sells the rest of lineup also even if they are worst then the competitor, its literally free marketing espeically for bandwagoners and they are everywhere

consumers arent the brightest you should know this because we all have our areas in products we buy where we are clueless

[Trixanity]

3 hours ago, Master Disaster said:

Oh you might be correct, it might have been the X800 that had the HL2 partnership.

From what I can tell it's both. I think HL2 had some significant delays so it'd make sense that it started on one and extended onto the next in the interim.

 

Yup, just checked. Original schedule was September 2003 which lines up with the release of the 9800 series. X800 came in like May 2004 and HL2 was finally released in November 2004 - over a year late.

[pas008]

8 hours ago, Master Disaster said:

Oh boy, so many young whipper snappers here.

 

The XT is a throwback to a legendary card from a long time ago...

 

https://www.techpowerup.com/gpu-specs/sapphire-9800-xt.b243

 

At a time when Nvidia were dominating the GPU market AMD dropped this thing, it performed similarly but was way cheaper and to top it off, they even got a Valve Half Life 2 partnership with 9800 Pro and XT models saying "Perfect for Half Life 2" right on the boxes. Some boxes even had Gordan and Alyx on them.

 

15 minutes ago, Trixanity said:

From what I can tell it's both. I think HL2 had some significant delays so it'd make sense that it started on one and extended onto the next in the interim.

 

Yup, just checked. Original schedule was September 2003 which lines up with the release of the 9800 series. X800 came in like May 2004 and HL2 was finally released in November 2004 - over a year late.

i remember my 9800 pro

with abit mobo

[mynameisjuan]

So AMD is just pulling a **** move with it's naming scheme now?

 

First with Ryzen taking Intel's next iteration now Nvidia's. It's just confusing people just getting into PC building.

[Master Disaster]

3 hours ago, pas008 said:

 

i remember my 9800 pro

with abit mobo

I had a 9800 Pro and I softmodded it to a 9800 XT, they really were mighty powerful cards for the time.

[Unixsystem]

8 hours ago, Neftex said:

thats my point, its the price/performance that decides. you dont need better performing halo product, you dont have to aim to be better than competitors top end. 

Why does it need to be one or the other? Obviously price/performance matters, but having a halo product has an effect as well.

 

The problem is that CPUs have so many different product lines and different workloads that it's hard to say what's objectively "best". I think what you're overlooking is the psychological effect that having the core advantage has. Even if the performance number is the same/slightly less, the type of consumer that is going to be easily swayed by appearance is also the one that is going to be most likely to be drawn in by big numbers, and when Ryzen 1 launched it had twice the cores as Intel's comparable CPUs, and pretty much the same thing happened with Threadripper. Things like IPC or memory latency are really hard to market, but core count is really easy. 

 

That said it won't work if the price/performance isn't also there, but I think a significant portion of Ryzen's success is likely due to the fact that they have been making Intel look like they were playing catchup since the day it was formally unveiled. Even if the performance doesn't give them an objective advantage, the massive push in terms of core count made them look like they had a halo product. 

[mr moose]

4 hours ago, Waffles13 said:

 the psychological effect that having the core advantage has

In many cases the Ryzens were at worst 10% behind Intel's best in games and on par or in front for most other end user situations.  The advantage Ryzen had (the reason it is a halo product) is clearly observable.  Meaning,  the reason people thought it was good wan't a psychological one, but an actual response to performance benchmarks and product reviews.

 

 

 

 

[mr moose]

16 hours ago, Trixanity said:

Minor anomalies? Nvidia have moved more units for like 20 years. With a single outlier where they moved less. That's the actual anomaly; that Radeons sold in bigger quantities.

Correct, so when AMD start to gain a little bit of market share that is an anomaly to what is being discussed.

16 hours ago, Trixanity said:

 

How often do you see Xeons mentioned in consumer marketing and the community at large?\

A fair but actually.  they even get recommended over the i7 when the end user isn't interested in overclocking.

16 hours ago, Trixanity said:

 

The reason AMD was in the red is that the CPU business is their main business and it was bleeding badly (started with the launch of Core 2 and went completely off the rails with Bulldozer).

It doesn't matter why really, the point is it was in the red and didn't turn around till Ryzen launched,   Fury did nothing, R290 and 390 did nothing, Exclusive sales to PS and Xbox did very little, Even having their GPU's in the mac pro did nothing.  Within 6 months of Ryzen and everything is climbing.

 

 

16 hours ago, Trixanity said:

Ryzen was a compelling product but only in a few instances of core count advantage. The most compelling aspect was good performance at a very affordable price (sometimes half the price as you moved up the stack). Why would you buy a more expensive Intel when you got more value at AMD? You needed parallel workloads for productivity to really get the most value. Otherwise if you want absolute performance it's Intel all the way. I even had to make that recommendation for people who wants high frame rate gaming. 

I don't understand this desire to try and dismiss Ryzen beyond being cheap.     It's like the idea that there were cases where it was better than Intel is an affront to existence. 

[Trixanity]

4 hours ago, mr moose said:

Correct, so when AMD start to gain a little bit of market share that is an anomaly to what is being discussed.

A fair but actually.  they even get recommended over the i7 when the end user isn't interested in overclocking.

It doesn't matter why really, the point is it was in the red and didn't turn around till Ryzen launched,   Fury did nothing, R290 and 390 did nothing, Exclusive sales to PS and Xbox did very little, Even having their GPU's in the mac pro did nothing.  Within 6 months of Ryzen and everything is climbing.

 

 

I don't understand this desire to try and dismiss Ryzen beyond being cheap.     It's like the idea that there were cases where it was better than Intel is an affront to existence. 

What is being discussed is that having the best halo product sells cards. When AMD has the best halo cards, they don't really sell more cards. That makes your argument invalid (in this case). You said that AMD needs a card to beat 2080 Ti to be of any interest. History tells us that AMD could launch a card 40% faster than 2080 Ti and it still wouldn't put much of a dent in the sales curve. The reason I can mention that number is because AMD has done that specifically in the past with no effect. So why would it be any different now? How is it going to change 20 years of history? 

 

Ok, let's just for the sake of argument say that Xeon gets recommended as an alternative to a Core product but tell me again how this relates to that a 56-core Xeon processor sells a Pentium Gold processor? Let's again assume for the sake of argument that it does. How does one go from looking at a Core i7 to another user saying you should get that Xeon equivalent that you can pick up for cheap because you know Xeon is awesome because you can get it in a 56-core variant? We have to jump through quite a few hoops to land that correlation. Best case scenario is strengthens the brand somewhat but do you really think that 56-core Xeon points your average consumer in that direction? Likewise I don't think a V100 sells a GTX 1060 in any significant way. A 2080 Ti does though.

 

It doesn't matter why? It doesn't matter that their CPU business is their main business, it's in the trash; and when it isn't, it makes them money? If their GPU business could have picked up the slack it wouldn't have been in the red. However we know how that went which is also why we know it matters.

 

You don't understand the desire because you're overselling what it was and argue that it's selling because it's better and/or there is some sort of halo product driving it forward. It didn't do what Core 2 did to AMD. It was competitive generally speaking but the real win was providing more cores for the workloads that can benefit for a cheaper price (while being dirt cheap to produce). Again if the premise is that price don't matter, you could pick up plenty of Xeons with massive core counts. They should at least have had 28 core processors at the time so it was certainly available if you wanted to pay for it. The fact that I've even had to recommend an Intel processor for gaming despite the higher price was a weird proposition at the time but we all know if you played at 1080p that performance wasn't quite there and if you wanted high frame rate it completely dropped off. Just as an example you could go anywhere from a couple of percent behind in FPS to 35%. That's why we're all talking about clock speed increases for the next generation. We wouldn't be talking about it if it wasn't an issue. There wouldn't be all these wild claims about 5 Ghz if it wasn't important. There are clear deficiencies. Ignoring that would be odd. However that doesn't mean I don't still recommend Ryzen whenever applicable because usually it's the better deal but again that wasn't the argument here. The argument was that absolute performance was good enough to classify it a halo product and/or that it was equivalent. Its main goal and main triumph was bringing many cores to the masses. It succeeded in that but it was behind on a per core performance which generally, for most users, is the most important metric and why many still paid the extra money to get that from Intel.

 

Basically I'm grounding the discussion instead of letting it drift skyward with all sorts of hypotheticals. We were talking AMD's lack of competitiveness in the high end and how changing it would solve their problems in market share. I have revealed that consumers give quite few fucks. Instead we got all sorts of Ryzen comparisons yet those sales weren't based on halo products either but just selling a gosh darn good value. So what can we learn from this? Analogies and comparisons don't automatically work despite the similarity of markets. I believe we're right about due for some more car analogies though; those are always fun.

[mr moose]

9 minutes ago, Trixanity said:

What is being discussed is that having the best halo product sells cards. When AMD has the best halo cards, they don't really sell more cards. That makes your argument invalid (in this case). You said that AMD needs a card to beat 2080 Ti to be of any interest. History tells us that AMD could launch a card 40% faster than 2080 Ti and it still wouldn't put much of a dent in the sales curve. The reason I can mention that number is because AMD has done that specifically in the past with no effect. So why would it be any different now? How is it going to change 20 years of history? 

 

Ok, let's just for the sake of argument say that Xeon gets recommended as an alternative to a Core product but tell me again how this relates to that a 56-core Xeon processor sells a Pentium Gold processor? Let's again assume for the sake of argument that it does. How does one go from looking at a Core i7 to another user saying you should get that Xeon equivalent that you can pick up for cheap because you know Xeon is awesome because you can get it in a 56-core variant? We have to jump through quite a few hoops to land that correlation. Best case scenario is strengthens the brand somewhat but do you really think that 56-core Xeon points your average consumer in that direction? Likewise I don't think a V100 sells a GTX 1060 in any significant way. A 2080 Ti does though.

 

It doesn't matter why? It doesn't matter that their CPU business is their main business, it's in the trash; and when it isn't, it makes them money? If their GPU business could have picked up the slack it wouldn't have been in the red. However we know how that went which is also why we know it matters.

 

You don't understand the desire because you're overselling what it was and argue that it's selling because it's better and/or there is some sort of halo product driving it forward. It didn't do what Core 2 did to AMD. It was competitive generally speaking but the real win was providing more cores for the workloads that can benefit for a cheaper price (while being dirt cheap to produce). Again if the premise is that price don't matter, you could pick up plenty of Xeons with massive core counts. They should at least have had 28 core processors at the time so it was certainly available if you wanted to pay for it. The fact that I've even had to recommend an Intel processor for gaming despite the higher price was a weird proposition at the time but we all know if you played at 1080p that performance wasn't quite there and if you wanted high frame rate it completely dropped off. Just as an example you could go anywhere from a couple of percent behind in FPS to 35%. That's why we're all talking about clock speed increases for the next generation. We wouldn't be talking about it if it wasn't an issue. There wouldn't be all these wild claims about 5 Ghz if it wasn't important. There are clear deficiencies. Ignoring that would be odd. However that doesn't mean I don't still recommend Ryzen whenever applicable because usually it's the better deal but again that wasn't the argument here. The argument was that absolute performance was good enough to classify it a halo product and/or that it was equivalent. Its main goal and main triumph was bringing many cores to the masses. It succeeded in that but it was behind on a per core performance which generally, for most users, is the most important metric and why many still paid the extra money to get that from Intel.

 

Basically I'm grounding the discussion instead of letting it drift skyward with all sorts of hypotheticals. We were talking AMD's lack of competitiveness in the high end and how changing it would solve their problems in market share. I have revealed that consumers give quite few fucks. Instead we got all sorts of Ryzen comparisons yet those sales weren't based on halo products either but just selling a gosh darn good value. So what can we learn from this? Analogies and comparisons don't automatically work despite the similarity of markets. I believe we're right about due for some more car analogies though; those are always fun.

Honestly mate, I think your way too invested in this.

[Trixanity]

3 minutes ago, mr moose said:

Honestly mate, I think your way too invested in this.

In other word: "I'm wrong"

[mr moose]

1 minute ago, Trixanity said:

In other word: "I'm wrong"

It's just not that complicated of a concept.   It's certainly not worth massive posts.  If your not willing to accept such a simple premise with due evidence then it's not worth continuing the discussion.  You are fixated on proving minute points that are neither here nor there.  Ryzen for all intents and purposes is a halo product, The closest thing in GPUs AMD has had is Vega 7 (which is too recent to see any real results in sales).  As has been pointed out by several people in this thread already, people look at who as the top when they buy from the bottom.  It's why cheap Samsung phones sell, its why the 1660 will sell, it's why Pentiums and Celerons sell.  It's something I have witnessed first hand in the industry. 

[Trixanity]

Just now, mr moose said:

It's just not that complicated of a concept.   It's certainly not worth massive posts.  If your not willing to accept such a simple premise with due evidence then it's not worth continuing the discussion.  You are fixated on proving minute points that are neither here nor there.  Ryzen for all intents and purposes is a halo product, The closest thing in GPUs AMD has had is Vega 7 (which is too recent to see any real results in sales).  As has been pointed out by several people in this thread already, people look at who as the top when they buy from the bottom.  It's why cheap Samsung phones sell, its why the 1660 will sell, it's why Pentiums and Celerons sell.  It's something I have witnessed first hand in the industry. 

Why are you so invested in this?

 

I thought we had established you were wrong and you're even saying as much now but just trying to get a few jabs in of the same arguments I've already addressed. Just accept defeat. You can't clutch this. In other words: omae wa mou shindeiru.

[mr moose]

1 minute ago, Trixanity said:

Why are you so invested in this?

 

 

I'm not,  hence why I did not read your wall of text,  if you need to say so much then you are over thinking the concept, clutching at minuscule attributes in order to dismiss a concept that obviously upsets you.  It's just the way humans are,  And probably a good thing too or else no one would be pushing technology further. 

[Trixanity]

2 minutes ago, mr moose said:

I'm not,  hence why I did not read your wall of text,  if you need to say so much then you are over thinking the concept, clutching at minuscule attributes in order to dismiss a concept that obviously upsets you.  It's just the way humans are,  And probably a good thing too or else no one would be pushing technology further. 

If you have so little to say it must mean there is no thought process behind and/or you can't explain your thoughts or concepts.

But yup, the fact you can't dispute what I've said tells me that you have nothing. Even gave you a TL;DR and you can't dispute that either.

Just throw in the towel or get someone else to do it.

[mr moose]

4 minutes ago, Trixanity said:

If you have so little to say it must mean there is no thought process behind and/or you can't explain your thoughts or concepts.

But yup, the fact you can't dispute what I've said tells me that you have nothing. Even gave you a TL;DR and you can't dispute that either.

Just throw in the towel or get someone else to do it.

you mean the bit that starts:

Quote

Basically I'm grounding the discussion instead of letting it drift skyward with all sorts of hypotheticals.

 

because that's not  TL:DR,  I actually read that sentence and decided it was just you saying you thought everyone else was being hypothetical and only you had the facts to ":ground" the discussion.  It doesn't warrant a response.

 

In fact you haven't even given a shred of evidence that having halo products doesn't elevate sales across the product stack.  Just your own opinions.

 

EDIT: in fact given you haven;t made a good argument in the last 10 or so posts unless you provide something substantial to counter what I and others have posted I am going to assuming you are just repeating yourself.   And I am not going to just repeat myself in response.

[leadeater]

26 minutes ago, Trixanity said:

290X was faster than Titan. I checked that ages ago and I just double checked. Either similar or better performance from 290X (and almost double the performance per dollar but I know you're not into that sort of thing).  

Kelper Titan was a bit slower, Kepler Titan Black (780 Ti) was very similar. Problem with the 290X was Nvidia's next generation, 900 series, came relatively too quickly to properly capitalize on the performance lead. Typical AMD lack of planning, market awareness and marketing.

 

290X was also stuck on reference blower designs for BLOODY AGES! So most people were waiting for custom AIB models which by then Geforce 900 was the next up coming big thing. And then mining, mining and more mining.

 

As a 290X owner I've always felt it was competing with the 900 series, the 980, then out classed by the 980 Ti/Titan X (Maxwell). All the while most had already gone out and brought 700 series before AMD released theirs and had no big reason to upgrade to a 290X with a garbage cooler or the 900 series until the 980 Ti.

 

What the market graph also shows is the effects of being late to the party, AMD performed better on the instances they released at the same time and badly if they did so after Nvidia.

[mr moose]

3 minutes ago, leadeater said:

What the market graph also shows is the effects of being late to the party, AMD performed better on the instances they released at the same time and badly if they did so after Nvidia.

 This is one of my points, maybe I am failing at language, how ever these little exceptions occur but don't disprove the condition.

[cj09beira]

32 minutes ago, leadeater said:

Kelper Titan was a bit slower, Kepler Titan Black (780 Ti) was very similar. Problem with the 290X was Nvidia's next generation, 900 series, came relatively too quickly to properly capitalize on the performance lead. Typical AMD lack of planning, market awareness and marketing.

 

290X was also stuck on reference blower designs for BLOODY AGES! So most people were waiting for custom AIB models which by then Geforce 900 was the next up coming big thing. And then mining, mining and more mining.

 

As a 290X owner I've always felt it was competing with the 900 series, the 980, then out classed by the 980 Ti/Titan X (Maxwell). All the while most had already gone out and brought 700 series before AMD released theirs and had no big reason to upgrade to a 290X with a garbage cooler or the 900 series until the 980 Ti.

 

What the market graph also shows is the effects of being late to the party, AMD performed better on the instances they released at the same time and badly if they did so after Nvidia.

amd really needs aibs since day 1, maybe they should set the gpu height and the screw holes, much sooner so that at least we get custom coolers at launch, really thats all amd cards need their pcbs are great. 

if i ever saw lisa that would be the one thing i would roast her for, and focusing on yields on choosing stock voltages.

[leadeater]

1 minute ago, cj09beira said:

amd really needs aibs since day 1, maybe they should set the gpu height and the screw holes, much sooner so that at least we get custom coolers at launch, really thats all amd cards need their pcbs are great. 

if i ever saw lisa that would be the one thing i would roast her for, and focusing on yields on choosing stock voltages.

Just ditch blower altogether, even a bad open air is vastly better. It feels like years ago someone screwed up a purchase order and brought 10 million blowers instead of 1 million lol

[Trixanity]

6 hours ago, leadeater said:

Kelper Titan was a bit slower, Kepler Titan Black (780 Ti) was very similar. Problem with the 290X was Nvidia's next generation, 900 series, came relatively too quickly to properly capitalize on the performance lead. Typical AMD lack of planning, market awareness and marketing.

 

290X was also stuck on reference blower designs for BLOODY AGES! So most people were waiting for custom AIB models which by then Geforce 900 was the next up coming big thing. And then mining, mining and more mining.

 

As a 290X owner I've always felt it was competing with the 900 series, the 980, then out classed by the 980 Ti/Titan X (Maxwell). All the while most had already gone out and brought 700 series before AMD released theirs and had no big reason to upgrade to a 290X with a garbage cooler or the 900 series until the 980 Ti.

 

What the market graph also shows is the effects of being late to the party, AMD performed better on the instances they released at the same time and badly if they did so after Nvidia.

That kinda ignores both sides of the equation. Both companies are essentially one-upping each other for a while except AMD doesn't gain much. HD 5000 is essentially the last card to make any kind of dent in the landscape. Otherwise we have to assume that Nvidia's timing ensured maximum numbers of lifecycle replacements and I find it odd that Nvidia would have timed it perfectly every single time. A year of a performance crown should be enough to get more than a few percentage points. Of course we can trace a lot of the issues back to the transition from X800 until HD4000/5000 and likewise the upward trend of Nvidia with the 8000 series going forward. Basically the shift of significant market share points hinges on one party doing very well and the other doing very poorly otherwise people stick to their guns and outside of something like Fermi then Nvidia has executed well enough to retain customers. The times AMD executed well wasn't to any big detriment of Nvidia nor did it last more than a year (usually).

Edited May 10, 2019 by wkdpaul
cleanup

[Drak3]

2 minutes ago, Trixanity said:

Both companies are essentially one-upping each other for a while except AMD doesn't gain much.

AMD doesn't gain much because Nvidia tends to release more powerful cards within a generation, and more recently, if AMD does release a better card, it's only better than Nvidia's offering for a short while. There is also the issue of AMD's drivers not being all that good for a long time, meaning that even if the hardware was better, performance wasn't when the cards were released.

[leadeater]

16 minutes ago, Trixanity said:

HD 5000 is essentially the last card to make any kind of dent in the landscape.

It was also pretty much the last card they released first that was also the performance pinnacle. HD 6000 series started as a re-brand then barely any improvement with the actual new architecture 6900 cards, I also owned a 6970. GTX 480 was earlier to market and just as fast as HD6970 (on avg and even closer on higher res) and GTX 580 was outright faster which came at similar time, so HD 6000 should have sold less than Nvidia.

 

7970 was an excellent card, also plagued by blowers on launch, then beaten by the 680 in most of the more popular titles.

 

After the 5870 AMD was either MIA or competing with blowers, doomed to failure.

[Trixanity]

51 minutes ago, leadeater said:

It was also pretty much the last card they released first that was also the performance pinnacle. HD 6000 series started as a re-brand then barely any improvement with the actual new architecture 6900 cards, I also owned a 6970. GTX 480 was earlier to market and just as fast as HD6970 (on avg and even closer on higher res) and GTX 580 was outright faster which came at similar time, so HD 6000 should have sold less than Nvidia.

 

7970 was an excellent card, also plagued by blowers on launch, then beaten by the 680 in most of the more popular titles.

 

After the 5870 AMD was either MIA or competing with blowers, doomed to failure.

I was under the impression that the argument was absolute performance sells cards; not price, packaging, power consumption or what have you. By that logic a blower cooler shouldn't matter in sales as long as it's the top dog in performance.

 

Or maybe it's just a contributing factor in a long list of factors but not the be all end all of sales strategery. That's all I'm saying and what I'm pointing out with the graph. There are in fact multiple graphs showing the same trend. That's why I'm saying a Navi card toppling  2080 Ti isn't gonna do much at all (hypothetically speaking of course). Not even if it's cheaper at the same time.

They've been there before.

 

Nvidia's 7nm would have to be very bad to lose any significant market share and realistically Nvidia has solid enough R&D for that not to happen. I think I've said it before but if not I'll say it again: AMD needs to be absolutely dominant across multiple years for them to on average move more units than Nvidia and to get to a point where we could even consider AMD at 50% (not accounting for Intel's plans and execution) market share. Realistically that's never gonna happen, not in any foreseeable future. They use the same foundry so there won't be any node advantage to rely on either.