Intel marketing slides - yawn.

Include the $$$ figures too.

Intel Rocket Lake Core i9-11900K is set to have a TDP/PL1 of 125 W and a PL2 turbo power limit of 250 W, with a turbo time of 56 seconds.

Let's also add this:

RTX3090 TDP 350 W Suggested PSU 750 W

This isn't how I imagined the future.

That isn't very impressive, honestly. Gaming performance is highly thread bound. A marginal victory in gaming might be enough to get gamers to buy the CPU, but it won't help their HEDT or enterprise segments at all.

GeForce GTX 590 (2011)- 365 W
GeForce GTX TITAN Z (2014) - 375 W
it looks they either used titan or x90 in a generaiton

I don't trust Intel's marketing slides...and even IF i did how much power and heat will these things put out? How much will they cost?
AMD can double the core count in ten seconds and be back on top in a jiffy if necessary.
Intel is floundering until they get on a modern process. No one wants another 14nm refresh, it's laughable.

just amd sells their doubled cores at the price of xeons

robertmo wrote on 2021-01-12, 17:15:

just amd sells their doubled cores at the price of xeons

If you are talking about Threadripper...then no they don't. They are roughly equivalent to Intel's HEDT sku's , at least as far as price is concerned.
They "smash" the hell out of Intel's HEDT platform in every other metric, speed, power usage, RAM channels, PCIe channels...etc.

AMD has their Epyc lineup to compete with Xeons....

robertmo wrote on 2021-01-12, 17:15:

just amd sells their doubled cores at the price of xeons

AMD abandoned the two ALUs for one FPU design with Zen. The weakness of Zen processors is the infinity fabric which connects each of the chiplets. When properly fed they are stronger than intel xeon in IPC and roughly similar in clock speed, but keeping them fed and using software that is NUMA aware, and in particular AMD thread scheduler aware, is on eof the challenges.

Largely, those challenges were solved in 2018 and 2019.

Zen 2 and Zen+ both struggled with intercore latency which really hurt gaming performance. Zen 3 has really fixed that in particular. Zen 3 is much faster in gaming compared to Zen 2 than it is faster in workstations compared to Zen 2 (although it is faster in both).

If Rocket Lake is faster on a per thread basis that is very impressive that intel was able to achieve that on a node that is about 30% less dense than TSMC 7nm. But that efficiency and performance comes at a cost. Rocket Lake will have a maximum of 8 cores and will only be for high performance desktops. Intel is still trying to bring Ice Lake Server to market on 10nm, which is a bit more dense than TSMC 7nm, but fraught with problems. I think even if Intel succeeds and brings 48 Ice Lake cores at 3.3ghz+ they will still have problems competing with AMD, but that is another topic.

With all of that said, pretty decent chance I might buy rocket lake myself 🤣. I'm watching it closely.

This reminds me of the 1Ghz PIII paper launches back in the day.

Intel has had trouble producing halo products at the quantities required lately.

https://www.anandtech.com/show/16341/intel-co … -intel-flagship

At this point someone like Intel has two choices.

They could draw the line at this exact intersection, regardless of performance. The performance and power figures would fall where they are, which in turn would affect the marketing strategy. The issue here is that the marketing strategy in-of-itself would directly affect how many of that product Intel tends to sell. Past sales performance is no guarantee of future success, and so this has to be managed.

The other choice is to draw the line at a point more aggressive, where Intel know it won’t be able to meet demand, but it will be able to leverage the increased performance processor in its marketing strategy and keep their premium product feeling premium. The problem here is if that line is drawn too aggressive – even those launch day performance figures look good, interest in the product will likely diminish if people can’t get hold of it, even with the higher performance level. This is important if system integrators that build machines directly to end-users can’t offer the flagship processor in their best systems.

Ultimately, this is what I think happened to Intel with the Core i9-10900K. The silicon quality level required to manufacturer the hardware was strict to provide a higher performance product, but too strict to be able to manufacturer a sufficient quantity to meet demand, especially for system integrators that rely on a steady source of good performance products. For all the plaudits Intel has received for eking out the 14nm process, the line for the 10900K was drawn too far, and the company wasn’t able to meet its own goals.

Furthermore, the only way Rocket Lake will overtake Ryzen 5000 on IPC is by pure clock speed. Even with their IPC improvements they are still going to be 3-4% behind AMD clock to clock. Now thats still a impressive jump, but if they can only maintain 5.3ghz on one core for 60 seconds at a time....im far less impressed. Maybe this will benefit in some games, but definitely not across the board. Productivity will remain largely in AMDs favor. Once you scale tye resolution past 1920x1080 a 4-5% lead is going to vanish.

I also think theybare having trouble hitting sustainable yields on these 14nm chips and that is why they reduced the core counts form 10 to 8. Intel has got to get a new process node working if they hope to keep up competition. Simple as that.

Why is it that when confronted with the choice of an AMD or Intel system, where the AMD system vastly outweighs the Intel system in the price/performance ratio, I still opt to spend more for the Intel system which is often two generations behind, performance-wise.

Is this a shortcoming on my part, or a reflection of my low opinion of AMD after all these decades of experience with their hardware?

Just some food for thought.

@mockingbird: AMD has some severe ups and downs. I was extremely happy with my Athlon XP and Athlon 64 back in the day. But their Phenom processors shit the bed so horribly, I didn't even think about AMD as a purchasing option until the goodwill towards the Ryzen 3's became too overwhelming to ignore.

I've never owned, nor even heard, of an Intel processor as objectively bad as AMD's misses have been. And going from the highs of the Athlon 64 to the lows of Phenom really shows you won't necessarily see those misses coming.

I was over the Intel vs AMD thing a couple of decades ago. Just use what I can get my grubby little old hands on...For nothing if possible 😉

Namrok wrote on 2021-01-12, 19:23:

@mockingbird: AMD has some severe ups and downs. I was extremely happy with my Athlon XP and Athlon 64 back in the day. But their Phenom processors shit the bed so horribly, I didn't even think about AMD as a purchasing option until the goodwill towards the Ryzen 3's became too overwhelming to ignore.

I've never owned, nor even heard, of an Intel processor as objectively bad as AMD's misses have been. And going from the highs of the Athlon 64 to the lows of Phenom really shows you won't necessarily see those misses coming.

Early P4's and later Pentium D chips were pretty damn bad objectively speaking.

Skylake and later generation core i chips haven't been "bad" objectively speaking. Just bad value for the most part.

And Intel's insistence on calling the last 3-4 14nm refreshes as "generation jumps" is completely laughable.

Increasing power limit and putting it on a new socket and maybe adding a few cores isn't really a generational improvement.

My opinon of course.

That being said....Bulldozer was almost unforgivably terrible.

Namrok wrote on 2021-01-12, 19:23:

@mockingbird: AMD has some severe ups and downs. I was extremely happy with my Athlon XP and Athlon 64 back in the day. But their Phenom processors shit the bed so horribly, I didn't even think about AMD as a purchasing option until the goodwill towards the Ryzen 3's became too overwhelming to ignore.

I've never owned, nor even heard, of an Intel processor as objectively bad as AMD's misses have been. And going from the highs of the Athlon 64 to the lows of Phenom really shows you won't necessarily see those misses coming.

Phenom wasn't really that bad. There were instruction problems on the phenom I, and it under delivered on performance compared to core2 but it was reasonably competitive, but phenom II was pretty smooth and it competed well with core2duo wolfdale processors at the time. The problem was that Intel released Nehalem at about the same time as AMD released phenom II and thuban, and nehalem made everything else that came before it look prehistoric. From 2008 to 2014 Intels tick-tock strategy was moving at light speed and nobody could keep up. AMD invested heavy in bulldozer which was a bigger mistake than netburst and that is how we got to 2018 with AMD stocks being the lowest they had ever been.

Zen is probably more successful than athlon64 at this point, however. Intel is in no danger of going bankgrupt, but they are losing huge mindshare and marketshare. Looking at Intel's roadmap going through 2024 I don't see things get much better. Alderlake could be great for gamers and mobile but I don't see it beating Zen 4 in servers (although too early to tell). It won't be until Redwood at least that intel could have a shot of dominating again.

What's with the childish "smashes" thing? - this is merely regurgitated marketing materials, showing approximately F-All.

mockingbird wrote on 2021-01-12, 19:05:

Why is it that when confronted with the choice of an AMD or Intel system, where the AMD system vastly outweighs the Intel system in the price/performance ratio, I still opt to spend more for the Intel system which is often two generations behind, performance-wise.

Is this a shortcoming on my part, or a reflection of my low opinion of AMD after all these decades of experience with their hardware?

Just some food for thought.

It's for your own thoughts, to determine how much of your choices are driven by objectivity, and how much by subjectivity. And judge yourself accordingly. Nobody else can know that for you.

I've had many and varied awesome systems for decades, based upon processors from both manufacturers. The merit of a given PC comes back to the skill and choices of the system builder. Not some silly "brand loyalty" thing.

Shreddoc wrote on 2021-01-12, 22:04:

I've had many and varied awesome systems for decades, based upon processors from both manufacturers. The merit of a given PC comes back to the skill and choices of the system builder. Not some silly "brand loyalty" thing.

I feel I should qualify what I said with something substantive in terms of explaining my dissatisfaction with AMD on the whole.

One example that comes up off the top of my head is back in 2018 when I ventured to reap the 'reward' of AMD's better performance. These were two simple desk/word processing machines I was selling. My main consideration for going with the Ryzen 2200G at the time was the incredible advantage it had at its price point to anything Intel had to offer.

But then after setting up the systems and going through the installation, I ended up again losing respect for AMD because of this:

"No AHCI driver exists for VEN_1022 DEV_43B7"

The most amazing part was the denial and apathy of the AMD team.

In retrospect, someone claimed you can force install their old AHCI driver on to it, but if that is indeed the case, all AMD had to do to save face was add 1 line to the AHCI .inf and get it WHQL'ed in the next release. I am 99% certain this problem still exists.

It's the little things, you see. Very nice silicon, but what is the value of good silicon if it's amateur hour in other departments.

Then there's Wendell who seems to release a new video about how "impressed" he is with the new generation of AMD hardware with regard to virtualization performance... Then you continue to watch the video and see that he's complaining at the end of lack of support for X on AMD's part, where he's forced to concede that it works 'perfect' with Intel. This is an ongoing saga now for several years on his channel. You mean to tell me that after all these years, AMD still doesn't have virtualization down pat to a fine art?

Next, try getting AMD part numbers which are OEM only... If I want a 35W TDP Intel part, I go to the large distributors and place a backorder for one piece. I might have to waith a month or two until it arrives, but I can get it.

Try obtaining a "GE" part from AMD. Nope, sorry, they only sell them to the privileged few OEMs.

I'll use AMD in the future, I have to. They have better products at better prices. But AMD isn't real competition to Intel. Intel allows AMD to survive but not thrive. One day Intel will open up it's enormous maw and swallow them up, and they'll disappear without a trace, and we'll be better off for it.

mockingbird wrote on 2021-01-12, 19:05:

Why is it that when confronted with the choice of an AMD or Intel system, where the AMD system vastly outweighs the Intel system in the price/performance ratio, I still opt to spend more for the Intel system which is often two generations behind, performance-wise.

Is this a shortcoming on my part, or a reflection of my low opinion of AMD after all these decades of experience with their hardware?

Just some food for thought.

I'm the same... and I was once a huge Cyrix fan (I know not AMD, just saying not Intel), then K6 and K6-2 fanboy, then later an Athlon (original) fan (well it needed a big fan didn't it haha), but since then I'll always opt for Intel. I think what sealed it for me were the HP Proliant 365 and 385 AMD Opteron servers we bought at work back in 2010 (maybe?) that we so awful at single threaded applications they didn't last long.