The quote I saw was 35% faster in FP for A77.25% faster in floating point compared to Cortex A76 - you do the math
ARM just announced a 64 core server CPU design under the N1 name, part of its Neoverse initiative.When ARM can release an 8/16/32 core chip (real 8 core, not 4 fast and 4 slow clores) at 5ghz
Why are we talking about ARM in this thread?
This is the main problem for making comparisons with x86, the software is simply not there yet for true real world comparisons, I'm excited to see how this unfolds going forward.When Adobe releases full version of Photoshop on iPad sometime later this year, perhaps we will see a better representation of real world performance on those ARM processors.
It would make sense to disrupt this space if there would be any growth in this market
that HPC wants to go Open Source, instead of CUDA
Intel mobile CPUs are currently on their 4th generation of Image Processign Unit, part of the CPU die. In terms of die area, the IPU takes as much as an entire Ice Lake core.The use case is important too - ARM cpu's are normally used with specialist chips in things like phones to do the video decode, photo enhancement, etc. This is different to PC's where the cpu basically does everything so it has to be a beast, but is a lot lot less efficient then the specialist hardware in phones at the jobs they are specialized for.
Ice Lake integrates Intel’s 4th generation Image Processing Unit (IPU). The IPU was first added with the dual-core Skylake mobile processors back in 2015. The 4th Gen IPU found in Ice Lake introduces a number of new enhancements. There is now support for 4K 30fps video capture support. Additionally, new hardware support for better de-noising has been implemented which can support 16 MP stills in low light conditions. The new IPU also has a concurrent image pipeline, supporting multiple different processing from the same camera streams, allowing a single camera to take the functionality of multiple sensors. A common example of that is devices with both IR and RGB cameras in the laptop bezel which can now be changed to a single camera. Intel says they are exposing more registers from the IPU to software in order to provide more flexibility for applications that make use of that for machine learning. It’s also worth noting that Intel integrated the MIPI interface onto the processor as well. Previously that was found on the chipset. The change significantly improves the latency, a required attribute needed for more advanced ML-specific applications. Some of those changes are designed to form the foundation for future generations of improvements.
That was an example, lots of different specialist hardware required, but it does highlight Intel's other problem. With Intel is you get a finished chip, you get no choice over what additional functionality it contains. With arm you get the design - and you can make that into whatever chip you want - it could have 128 custom AI engines, and a couple of ARM cores for example. Or some 8k image encode/decode unit that can do 120fps and a few ARM cores if you wanted.Intel mobile CPUs are currently on their 4th generation of Image Processign Unit, part of the CPU die. In terms of die area, the IPU takes as much as an entire Ice Lake core.
The quote I saw was 35% faster in FP for A77.
Not sure. It's almost like a mod merged the wrong threads together.
A MD
R egains
M arket share
-
Nope, VKHR is a working project, you can also compile it from a source and try it by yourself - GitHub - CaffeineViking/vkhrHmmm, according to a guy on reddit, the VKHR thing is more of a research paper than a successor to TressFX, possibly might even end up as part of it.
Something to do with using volume based raymarching to increase the efficiency of hair rendering at a distance, which might make it easier to have several hairy characters on screen at once possibly.
Anand's own benchmarks certainly do not agree on ARM being a billion million times worse than x86, as some here like to insinuate:
https://www.anandtech.com/show/13786/snapdragon-855-performance-preview/2
https://www.anandtech.com/show/13405/intel-10nm-cannon-lake-and-core-i3-8121u-deep-dive-review/8
Here is a 2.8ghz A76 compared to 2.2ghz Kaby Lake/Cannon Lake on a single-threaded benchmark, it's not 2012 anymore ARM cores have come a long, long way.
Good information, thank you. As I said, definitely not well-versed on this. I have heard ARM also has micro ops.You can't determine execution time on any modern CPU by how many opcodes are used. On a modern CPU, opcodes are broken down into micro-ops and pipelined, so multiple opcodes may be executed in parallel (or speculatively, in the case of branches, which we recently found out causes a lot of potential security issues). And on x86, some of the older and less frequently used opcodes take multiple micro-ops.
In your example above, the rep cmps instruction is an antiquated instruction that most compilers don't generate any more because it is slower on a modern x86 CPU than just using regular comparison operations in a loop. For instance, on Skylake, rep cmps consists of 8 or more micro-ops per iteration, and on Ryzen, it's 9 micro-ops per iteration.
little off topic, but I wouldn't bring stock up in here, nvidia was MASSIVELY overpriced at nearing $300 at its peek, whilst AMD's previous peek of $34 (up from $8-10, right before falling back to a stable $18-20) was much more realistic, only a little too early.I did a quick google search and from what I can see Nvidia had higher revenue in 2017 than AMD in 2018.
On the other hand though the stock value of Nvidia has dropped over the last year whereas AMD has increased. Not only that, AMD's value has gone up since Zen got introduced and keeps climbing.
The one big drop they both share is when sales dropped due to crypto mining collapsing. However, AMD recovered from that stock value drop whereas Nvidia didn't.
So I think it's questionable whether or not Nvidia could get into a price war with AMD and then get out of that unscathed, since investors certainly don't want less revenue at this point (or any point). AMD can soak up losses elsewhere in its product portfolio whereas Nvidia really can't to the same degree.
what I say to that: first and foremost Adobe should optimize proper multithreading in their producuts that have made them rich in the past 15 years... both PS and Premiere are unoptimized garbabe right now, that don't run almost any better on a modern CPU than on a 7700K...This is the main problem for making comparisons with x86, the software is simply not there yet for true real world comparisons, I'm excited to see how this unfolds going forward.
Moving on to AMD graphics specific news, I found something on Phoronix about a new Vulkan based TressFX successor called VKHR, presumably it will be part of the FidelityFX initiative as it matures:
Link here.
Thread creep on steroids. Nice recursive acronym, by the way.
So, anyone know if Navi has any tiny ARM cores in there somewhere? Like TrustZone only . . . I dunno?
So much for Navi, even discussing ARM minutia is better.
Also I wouldn't look at stock price at all, AMD market cap is $33.0B, NVIDIA is $93.3B, Intel is $211.1B. That tells more of the story.little off topic, but I wouldn't bring stock up in here, nvidia was MASSIVELY overpriced at nearing $300 at its peek, whilst AMD's previous peek of $34 (up from $8-10, right before falling back to a stable $18-20) was much more realistic, only a little too early.
No it doesn't. Uber commands a market cap of $60B yet it loses billions of dollars and is arguably replaceable almost overnight due to no real technology advantage. Don't trust stock markets to value companies in any logical way.Also I wouldn't look at stock price at all, AMD market cap is $33.0B, NVIDIA is $93.3B, Intel is $211.1B. That tells more of the story.
Everybody deserves one free chance. Try again.This just in! AMD partnership with Uber! Buy 50th Anniversary Edition, have it delivered straight to you by UBER!
AMD is placing Navi as VEGA 56/64 replacement.