confirmed: AMD's big Navi launch to disrupt 4K gaming

Edgar said:
Remember when high end was $500. Those were the good ol' days.
Click to expand...
Back then the price per transistor was decreasing faster than the rate at which transistor density was increasing. Price per transistor has been pretty much flat for the past 4-5 years. If you want faster video cards, you're going to have to pay more for them. It's just the current reality of nanoscale technology.
ManofGod said:
When was that, 2005?
Click to expand...
GTX 680 and HD 7970 GHz are the last two I remember launching at $500, which was 2012.
 
GoodBoy said:
LOL. Stupid article, must be AMD fanboi's.

Steam has 90 million users. Using December 2019's Steam Hardware Survey numbers:
0.50% 2080Ti = 450,000 cards. That's over a half bilion dollars in sales btw.
1.34% 1080Ti = 1,206,000 cards.
The much beloved (by the aforementioned fanbois) AMD 5700Xt is at 0.22% = 198,000 cards. Sales of about 79 million using $400 price.
Top card with a crazy 20.91%! GTX1060 = 18,819,000 cards!
Click to expand...

I think you missed the point. It's about how most PC gamers will not or can not buy a RTX 2080 Ti but you rambled on about profits for some reason or other and BTW ... it's "boys" not "bois" o_O
 
Zarathustra[H] said:
You take a large enough random sample and it predicts the population closely.

Statistics fundamentals ftw.
Click to expand...
Not so sure in this case. Statistics fundamentals require proper statistics conduct. Since Steam's methodology isn't public, and some received the questionnaire multiple times and some never, I don't know if it is sufficiently random to extrapolate hard numbers. Trends and general picture, sure.
 
I read what AMD and nVidia are supposedly launching in the coming months and I don't see how either of them will "disrupt" 4K gaming, release a card capable of 4K gaming with full eye candy and playable framerates sure but that is hardly a disruption that is what we are asking for..... I see both of them though releasing solid 1080p cards in the $200 range and 1440p cards at sub $500 though and that will just be sweet. 1080p is still king, 1440p is the up and comming while 4K is still for the few. I am interested to see what they can get the entry costs of VR down too though, the headsets are coming down in price but are still significant so any means to decrease the overhead on that is welcomed to me.
 
GoodBoy said:
It's not 100% of the user base, never said it was. It is enough to be statistically accurate within the standard deviation. You can't "lie" on your hardware survey, that is automatically pulled if you opt in to the survey. Believe what you want. I'm sure AMD is king of the flat Earth in some reality...
Click to expand...

Mega6 said:
"automatically pulled" implies no user interaction. Misrepresentation abounds.
Click to expand...
 
Meeho said:
Not so sure in this case. Statistics fundamentals require proper statistics conduct. Since Steam's methodology isn't public, and some received the questionnaire multiple times and some never, I don't know if it is sufficiently random to extrapolate hard numbers. Trends and general picture, sure.
Click to expand...


That is true, there is a lot we don't know, but overall it works like this:


The accuracy by which a statistical sample like this is predictive of the overall population is based on the binomial distribution.


Code: 
Descriptive Statistics
N             Event    Sample p              95% CI for p
200         1              0.005000              (0.000127, 0.027542)

In the test above (with made up numbers) we have taken a sample of 200 out of the overall population. 1 of those 200 had a 2080ti.

In our sample, this is exactly 0.5% (Sample p) but at a 95% confidence level we can see that it predicts that in the overall population, the true occurrence rate is somewhere between 0.127% and 2.7542%, as long as it is a true random sample. If there is anything that skews the sample, these numbers don't hold up.

Essentially what we are saying here is that based on the sampling above, I am 95% sure that the true population value lies somewhere between 0.127% and 2.7542%

Code: 
Descriptive Statistics
N             Event    Sample p              95% CI for p
1000       5              0.005000              (0.001625, 0.011629)

Now, if we double the sample size to 1,000, and find that 5 in the sample are 2080ti's, our sample still has exactly 0.5% 2080ti's, but our prediction of the population as a whole has tightened. Now we know that it is somewhere in between 0.1625% and 1.1629%

Thus as we can see below by increasing the sample size we get a better and better prediction of the population...

Code: 
Descriptive Statistics
N             Event    Sample p              95% CI for p
2000       10           0.005000              (0.002400, 0.009176)

Descriptive Statistics
N             Event    Sample p              95% CI for p
5000       25           0.005000              (0.003238, 0.007372)

Descriptive Statistics
N             Event    Sample p              95% CI for p
10000    50           0.005000              (0.003713, 0.006587)

Descriptive Statistics
N             Event    Sample p              95% CI for p
20000    100         0.005000              (0.004070, 0.006078)

Descriptive Statistics
N             Event    Sample p              95% CI for p
50000    250         0.005000              (0.004401, 0.005658)

Descriptive Statistics
N             Event    Sample p              95% CI for p
100000  500         0.005000              (0.004572, 0.005457)

Descriptive Statistics
N             Event    Sample p              95% CI for p
1000000                5000       0.005000              (0.004863, 0.005140)

...until a sample size of 100,000 contains 500 2080ti's we know the true population is somewhere between 0.4572% and 0.5457%

A sample of 1,000,000 containing 5,000 2080ti's tells us that at the 95% confidence level we know that the true population is somewhere between 0.4863% and 0.514%

So, the larger sample we take, the more accurately we can predict what the overall population looks like. If we want a higher level of confidence we have the correct range, we could - for instance - raise the confidence for the calculation to 99%, but what happens then is that for the same sample size, our range of prediction on the overall population widens, because now we want to be 99% sure that the true population number is somewhere in that confidence interval.

I don't know the sample sizes Steam uses for each months statistics, so I can't tell you exactly how accurate they are, but considering their huge user base, and relative ease of collecting data, I bet the sample size is pretty large. Honestly, I don't understand why they don't just collect all data from all users, and make it even more accurate...


Samplings like these can be wrong though, and that is if they are not truly random samples, but instead something else is affecting the sample.

In the case of Steam, an example of something that could have an effect is that they ask for your consent before the system specs are submitted. (We have all seen that window). I don't think it is a stretch to guess that hardware enthusiasts who have invested money and time into their system are going to be much more likely to be eager to have their system counted, than the average user, who might just decline. This could skew the data, and make the sample non-random, resulting in numbers in the data skewing towards higher end systems.

Or the opposite could be true. Maybe privacy enthusiasts are more likely to also be hardware enthusiasts, and because of this, when they decline having their data submitted, high end systems are underrepresented.

Who knows which it is, but it is probably fairly certain that trends in who opts to submit, and who declines has some sort of impact on the overall data collection, and thus an impact on the quality of the statistical prediction.
 
GoodBoy said:
It's not 100% of the user base, never said it was. It is enough to be statistically accurate within the standard deviation. You can't "lie" on your hardware survey, that is automatically pulled if you opt in to the survey. Believe what you want. I'm sure AMD is king of the flat Earth in some reality...
Click to expand...



See my post above. Who chooses to opt in and who doesn't likely has some sort of impact on the randomness of the sample.

Whether or not it is large enough to have a significant impact on the results is impossible to know without further study and more detail than Valve provides.
 
ChadD said:
4K is the only resolution that really needs power today or for the next number of years.

I picked up a 5700xt and am playing on a ultrawide so a few more pixels to push over 1080p... and very little isn't running over 100fps at ultra settings.
Click to expand...

I have a 2080TI and even at 1440p there is a ton of stuff that won't run at over 100fps. At Ultra settings RDR2 is like 50-60fps. Hell Let Loose is 80-90fps. Then add in Control, Shadow of the Tomb Raider, Division 2, Metro Exodus.


The 5700XT can't even hold 60fps on AC:O at 1440p so the statement "4k is the only resolution that really needs power today or for the next number of years" is quite false when even the fastest card available can struggle at 1440p.

assassins-creed-odyssey-2560-1440.png
 
Apparently there's no controls to opt in any more. Since I installed Linux a couple of weeks ago I want Steam to count my Penguinista account. :geek:
 
Derangel said:
There is no such thing as "reasonable" and "unreasonable" prices. It's all about what the market will or will not support.
Click to expand...

Respectfully disagree. There is a reasonable price, however the masses arent going to know what that is and marketing can inflate value. It is up to people that are more knowledgeable than me to push back on these inflated prices. $1200 is not reasonable. $699 is reasonable, $799 is status quo, $1200 is a joke.
 
Zarathustra[H] said:
I don't think it is a stretch to guess that hardware enthusiasts who have invested money and time into their system are going to be much more likely to be eager to have their system counted, than the average user, who might just decline. This could skew the data, and make the sample non-random, resulting in numbers in the data skewing towards higher end systems.

Or the opposite could be true. Maybe privacy enthusiasts are more likely to also be hardware enthusiasts, and because of this, when they decline having their data submitted, high end systems are underrepresented.

Who knows which it is, but it is probably fairly certain that trends in who opts to submit, and who declines has some sort of impact on the overall data collection, and thus an impact on the quality of the statistical prediction.
Click to expand...
Statistics 101 for today? :)

Anyway, the above was exactly my thought process. First I was thinking that hardware enthusiast were more likely to participate, but then again more likely to care about privacy. I'm not certain which would prevail, but probably a smaller fraction would care enough about it to decline the survey.

What also made me question the validity of the survey to draw specific conclusions was the relatively high and constant rise of the 1060's share. Not sure what to make of that. OEM systems?
 
Meeho said:
Statistics 101 for today? :)

Anyway, the above was exactly my thought process. First I was thinking that hardware enthusiast were more likely to participate, but then again more likely to care about privacy. I'm not certain which would prevail, but probably a smaller fraction would care enough about it to decline the survey.

What also made me question the validity of the survey to draw specific conclusions was the relatively high and constant rise of the 1060's share. Not sure what to make of that. OEM systems?
Click to expand...

I think there are a ton of cheapish gaming laptops sold with 1060's, but I could be wrong. I don't normally follow that space.
 
Meeho said:
What also made me question the validity of the survey to draw specific conclusions was the relatively high and constant rise of the 1060's share. Not sure what to make of that. OEM systems?
Click to expand...

Could be OEM systems, could also be scammers/fraudsters getting each of their 100 accounts into the survey and thus being over-represented.
 
ChadD said:
Games have stagnated... the only real disruption that is possible at this point. Is if more people really buy into 4k. At this point where we need the disruption is in the Monitor space. lol Facts are unavoidable according to Valve there are more people playing at 1024x768 and 1280x1024 is about = to the number of people playing at 4k. Almost 3 quarters of gamers are playing at 1080p... and the ultrawide resolutions are almost in the double digits (9% for the ultra wide resolutions vs 2% for 4k).
Click to expand...

Not really, but yes games have slowed down a bit in graphical leaps. And so has the GPU cycle length to match it. The RTX 2070 struggles at keeping 60 frame rates at 2560x1440 in some games at max settings. Assassin's Creed Odyssey and Metro Exodus are examples. I am sure some of the upcoming games such as Cyberpunk, possibly the new Watch Dogs and some other titles will knock that down a bit further. For people who run this moderate resolution who enjoy high frame rates (60 or above) the 2070 level isn't going to cut it very well in the coming months. So you can easily argue that 1440 or 1440 ultra wide are due for a new generation as well.
 
Thunderdolt said:
Could be OEM systems, could also be scammers/fraudsters getting each of their 100 accounts into the survey and thus being over-represented.
Click to expand...

You can't "get yourself into the survey". It's a random selection of users, and if you get the popup, you can choose to participate or not. I've seen it all of twice in what 16+ years on Steam? (joined on the beta launch).
People are not skewing the results running 100 accounts thru the survey...
 
Last edited:
Re above quote, I'm willing to bet 1024x768 and 1280x1024 resolutions in the steam survey are drastically over represented due to counter-strike players (a valve game) weighing in on these surveys. Almost everyone who plays that game somewhat seriously uses a 4:3 resolution, typically to stretch it and make heads appear larger. So they're doing that as a style choice rather than a hardware limitation.

I think 4k would have taken off way more on the PC by now and was surely expected to, but adoption was stifled by the realization of high refresh monitors. Outside of slow moving adventure games like the witcher 3, i would rather (and do) play at 1080p 144hz over 4k 60hz. Mouse movement in first person just feels horrible bordering on unbearable at 60 hz to me now.

4k is beautiful but alas, i want 100+ fps at 4k before i really adopt it :p. I have zero expectation of big navi or rtx 3000 giving us high fps @ 4k unless sli/crossfire is reinvigorated. 2 cards would definitely do it.
 
Everyone is in here talking about resolution and I'm just sitting here wondering when something other than resolution will push hardware. I sure do miss the old days.
 
If AMD really wants to disrupt the highend market.

Make 2 top end SKUs - One without RT and one with RT. Sell the one without RT for $200-250 less. If the drivers are stable it will probably sell like hot cakes if the price isn't crazy. It will be in its own segment of the market.



(and we can always hope we can unlock hardware RT through BIOS flash hehe)
 
Zarathustra[H] said:
See my post above. Who chooses to opt in and who doesn't likely has some sort of impact on the randomness of the sample.

Whether or not it is large enough to have a significant impact on the results is impossible to know without further study and more detail than Valve provides.
Click to expand...

Nice refresher on sampling margins of error.

Because of the Opt In part, the Steam survey is not random.

1. You must be a gamer.
2. You must use Steam.
3. You allow Steam to "sample" your rig. (<--- This would be biased to those "proud" of their rigs, IMO.)

I have no idea how the Gamer/Steam/Opt-in subset reflects the overall PC population, or any part of the population not in that subset (e.g., Gamer/Steam, but not Opt-in).

I would think it is skewed to show more powerful gaming rigs as having a higher proportion than in actuality.

But I'd still like Big Navi to disrupt the $1200 price point. ;)
 
I'm most interested in Y4k gaming. Nonvirtual reality is where things are headed.

And since everything is chrome in the future, ray tracing will be a given.
 
GhostCow said:
Everyone is in here talking about resolution and I'm just sitting here wondering when something other than resolution will push hardware. I sure do miss the old days.
Click to expand...

Well it's resolution and FPS if you're looking to improve the current play experience; the higher you have of the former, the more difficult it is to produce the latter.

Ray tracing seems to be where the GPU vendors would like to head next, but we're hardly at the stage where the tech is ready to just take over.
 
c3k said:
I would think it is skewed to show more powerful gaming rigs as having a higher proportion than in actuality.
Click to expand...

If that is the case, then the numbers should indicate this. Do they? No, not really. The #1 GPU in the survey, at 15.8%, is the GTX 1060. That was one step up from the bottom when it was released back in July 2016. That card has nearly double the presence of the second place card, the 1050 Ti - which is also an aging card that was entry level when it was new. In fact, the most popular high end card on Steam is the 1080 Ti at only 1.53%. On the CPU front, 52% are running 4-core CPUs. Again, that's skewed to the high end? I don't think so.

If we assume for a moment that you are correct and the results ARE skewed towards the high end, what does that say about the state of gaming today? A three and a half year old entry level card represents the high end? That's kind of scary to think about.
 
Aix. said:
Well it's resolution and FPS if you're looking to improve the current play experience; the higher you have of the former, the more difficult it is to produce the latter.

Ray tracing seems to be where the GPU vendors would like to head next, but we're hardly at the stage where the tech is ready to just take over.
Click to expand...

Personally I would rather see better physics and most of the things currently being made with textures being made with polygons instead. Ray tracing is neat but that's not even using regular hardware. It has it's own dedicated hardware. At least with Nvidia's implementation.

We need a new Crysis. I'm tired of being held back by consoles.
 
Zarathustra[H] said:
That is true, there is a lot we don't know, but overall it works like this:


The accuracy by which a statistical sample like this is predictive of the overall population is based on the binomial distribution.


Code: 
Descriptive Statistics
N             Event    Sample p              95% CI for p
200         1              0.005000              (0.000127, 0.027542)

In the test above (with made up numbers) we have taken a sample of 200 out of the overall population. 1 of those 200 had a 2080ti.

In our sample, this is exactly 0.5% (Sample p) but at a 95% confidence level we can see that it predicts that in the overall population, the true occurrence rate is somewhere between 0.127% and 2.7542%, as long as it is a true random sample. If there is anything that skews the sample, these numbers don't hold up.

Essentially what we are saying here is that based on the sampling above, I am 95% sure that the true population value lies somewhere between 0.127% and 2.7542%

Code: 
Descriptive Statistics
N             Event    Sample p              95% CI for p
1000       5              0.005000              (0.001625, 0.011629)

Now, if we double the sample size to 1,000, and find that 5 in the sample are 2080ti's, our sample still has exactly 0.5% 2080ti's, but our prediction of the population as a whole has tightened. Now we know that it is somewhere in between 0.1625% and 1.1629%

Thus as we can see below by increasing the sample size we get a better and better prediction of the population...

Code: 
Descriptive Statistics
N             Event    Sample p              95% CI for p
2000       10           0.005000              (0.002400, 0.009176)

Descriptive Statistics
N             Event    Sample p              95% CI for p
5000       25           0.005000              (0.003238, 0.007372)

Descriptive Statistics
N             Event    Sample p              95% CI for p
10000    50           0.005000              (0.003713, 0.006587)

Descriptive Statistics
N             Event    Sample p              95% CI for p
20000    100         0.005000              (0.004070, 0.006078)

Descriptive Statistics
N             Event    Sample p              95% CI for p
50000    250         0.005000              (0.004401, 0.005658)

Descriptive Statistics
N             Event    Sample p              95% CI for p
100000  500         0.005000              (0.004572, 0.005457)

Descriptive Statistics
N             Event    Sample p              95% CI for p
1000000                5000       0.005000              (0.004863, 0.005140)

...until a sample size of 100,000 contains 500 2080ti's we know the true population is somewhere between 0.4572% and 0.5457%

A sample of 1,000,000 containing 5,000 2080ti's tells us that at the 95% confidence level we know that the true population is somewhere between 0.4863% and 0.514%

So, the larger sample we take, the more accurately we can predict what the overall population looks like. If we want a higher level of confidence we have the correct range, we could - for instance - raise the confidence for the calculation to 99%, but what happens then is that for the same sample size, our range of prediction on the overall population widens, because now we want to be 99% sure that the true population number is somewhere in that confidence interval.

I don't know the sample sizes Steam uses for each months statistics, so I can't tell you exactly how accurate they are, but considering their huge user base, and relative ease of collecting data, I bet the sample size is pretty large. Honestly, I don't understand why they don't just collect all data from all users, and make it even more accurate...


Samplings like these can be wrong though, and that is if they are not truly random samples, but instead something else is affecting the sample.

In the case of Steam, an example of something that could have an effect is that they ask for your consent before the system specs are submitted. (We have all seen that window). I don't think it is a stretch to guess that hardware enthusiasts who have invested money and time into their system are going to be much more likely to be eager to have their system counted, than the average user, who might just decline. This could skew the data, and make the sample non-random, resulting in numbers in the data skewing towards higher end systems.

Or the opposite could be true. Maybe privacy enthusiasts are more likely to also be hardware enthusiasts, and because of this, when they decline having their data submitted, high end systems are underrepresented.

Who knows which it is, but it is probably fairly certain that trends in who opts to submit, and who declines has some sort of impact on the overall data collection, and thus an impact on the quality of the statistical prediction.
Click to expand...
I have never seen that survey on any of my 5 desktops.... On 4 different steam accounts. So I don't know how random it really is, lol. Chances are I'd of seen one by now. I agree, I don't know why they don't just collect all data about CPU, ram and GPU... This would make their #'s more meaningful and less chance of something skewing results. (Puts on tin foil hat) Maybe it's because I have all AMD GPUs ;).
 
Ready4Dis said:
I have never seen that survey on any of my 5 desktops.... On 4 different steam accounts. So I don't know how random it really is, lol. Chances are I'd of seen one by now. I agree, I don't know why they don't just collect all data about CPU, ram and GPU... This would make their #'s more meaningful and less chance of something skewing results. (Puts on tin foil hat) Maybe it's because I have all AMD GPUs ;).
Click to expand...

I haven't seen one in a long time, but I used to see them fairly regularly.

My guess is that they have kept their sample size fixed, but the total number of users has expanded, meaning that over time any give individual is less likely to be picked.
 
  • Like
Reactions: N4CR
like this
Zarathustra[H] said:
I haven't seen one in a long time, but I used to see them fairly regularly.

My guess is that they have kept their sample size fixed, but the total number of users has expanded, meaning that over time any give individual is less likely to be picked.
Click to expand...
Ive had a steam account for like 7 years... Have gotten asked exactly 0 times...
 
Ready4Dis said:
Ive had a steam account for like 7 years... Have gotten asked exactly 0 times...
Click to expand...

Weird.

I guess that's how random sampling works though. Somewhere out there on the extreme tails of the distribution there is a non-zero probability of someone never being sampled. That just happens to be you I guess :p

That said, this year will be my 17th year on Steam. I signed up to pre-order Half-Life 2 and Cointer-Strike:Source.

I can't rememeber when the surveys started to become less frequent for me. Could very well have been more than 7 years ago. The last 7 years have really gone by in a blur.
 
Last edited:
After reading the thread at [H]ardform, AMD has decided to sell their entire GPU technology and division to Intel.

"People were going to start killing each other. We decided to do the right thing."

Intel is planning their first Navinside (aka RickiLake) card which is supposed to dominate the world of SVGA gaming. It's going to be built at TI's state of the art 65nm fab and will feature user upgradeable GPU memory using high speed Compact Flash. Early pics clearly showed the memory slots, plus one more that looked a lot like a GameBoy cartridge slot.

"Once you've seen Kirby in full SVGA, you'll never go back! We're going to change the world."
 
Zarathustra[H] said:
You take a large enough random sample and it predicts the population closely.

Statistics fundamentals ftw.
Click to expand...

No way, everyone knows that a very large random sample of gamers is always biased toward nvidia.

Apparently AMD users are more likely to decline Steam surveys for some reason.
 
Ready4Dis said:
I have never seen that survey on any of my 5 desktops.... On 4 different steam accounts. So I don't know how random it really is, lol. Chances are I'd of seen one by now. I agree, I don't know why they don't just collect all data about CPU, ram and GPU... This would make their #'s more meaningful and less chance of something skewing results. (Puts on tin foil hat) Maybe it's because I have all AMD GPUs ;).
Click to expand...

That’s not how randomness works. Each year you have an x (amount of surveys they send) out of y (total number of Stream accounts) chance. The y increases every single year, giving you less of a chance as time goes on. Just because you don’t get it one year does not mean your chances to get it go up the next.
 
I've certainly gotten it less often as time goes on. Used to be every year or two for me. I think the one I got this year was the first one in 3+ years
 
I've been asked about half a dozen times over the last 2 years. Declined each time.
 
Just a thought:

( This is my own speculation & not a rumor or news )

What if Big Navi released in 2020, does NOT have Ray Tracing ?

The die space saved will allow AMD to compete better with 2080ti at more affordable price, just like initial zen CPUs didn't have any igp which meant AMD could pack them with more CPU cores

AMD claimed to have 60% uplift going from integrated Vega in 12nm Zen+ to integrated Vega in Zen2 7nm Renoir APU

Maybe they could just apply the lessons learnt to RDNA 1, refresh it double the die size without doubling the power budget & release a RX 5900 that performs better than RTX 2080 ti for approx $700



https://www.anandtech.com/show/1544...fresh-and-rdna2-both-in-2020-according-to-amd
 
Marees said:
Just a thought:

( This is my own speculation & not a rumor or news )

What if Big Navi released in 2020, does NOT have Ray Tracing ?

The die space saved will allow AMD to compete better with 2080ti at more affordable price, just like initial zen CPUs didn't have any igp which meant AMD could pack them with more CPU cores

AMD claimed to have 60% uplift going from integrated Vega in 12nm Zen+ to integrated Vega in Zen2 7nm Renoir APU

Maybe they could just apply the lessons learnt to RDNA 1, refresh it double the die size without doubling the power budget & release a RX 5900 that performs better than RTX 2080 ti for approx $700



https://www.anandtech.com/show/1544...fresh-and-rdna2-both-in-2020-according-to-amd
Click to expand...

Is there a market for a 4K 60fps ultra $700 card that will never do RT?

E-sports, probably.
 
Minor point first:


Marees said:
AMD claimed to have 60% uplift going from integrated Vega in 12nm Zen+ to integrated Vega in Zen2 7nm Renoir APU
Click to expand...

AMD is already on 7nm with it's Navi parts. There is no 60% process change uplift to grab, and this is really only in Mobile where you are really keepting wattage tamped down.

There isn't likely going to be much process benefit for Navi going 7nm+. NVidia is definitely going to get a much bigger boost going from 12nm to 7nm (likely straight to 7nm+ as it looks to be arriving with big Navi).

Just a thought:

( This is my own speculation & not a rumor or news )

What if Big Navi released in 2020, does NOT have Ray Tracing ?

The die space saved will allow AMD to compete better with 2080ti at more affordable price, just like initial zen CPUs didn't have any igp which meant AMD could pack them with more CPU cores

https://www.anandtech.com/show/1544...fresh-and-rdna2-both-in-2020-according-to-amd
Click to expand...


Of all the "What ifs", that is the least likely.

I have been saying all along, that the main reason "Big Navi" was so much later than the rest, is because at the top end, Ray Tracing will be a must have. That point has arrived.

I expect the main difference between RDNA1 and RDNA2 will be Ray Tracing HW.

Bottom line: There is essentially ZERO chance of Big Navi arriving without Ray Tracing HW.
 
Auer said:
4 or 5 times here, but only in the last 2-3 years
Click to expand...
Yeah, that's what I'm saying, 0 in over 7 years (can't remember exactly), and I have had 4 accounts (mine + 3 kids) for a few years now and 0 on all of them. Seems odd as others have seen it a few times before. Guess I have no reason to buy lotto tickets with my luck!
 
You must log in or register to reply here.
Back
Top