On JavaScript benchmarks like octane and antutu html5, A11 clearly gave intel i5, and even i7s a run for the money. I tested it personally on my laptop which runs a skylake i5u
On JavaScript benchmarks like octane and antutu html5, A11 clearly gave intel i5, and even i7s a run for the money. I tested it personally on my laptop which runs a skylake i5uagain, you need to know and understand exactly what geekbench tests and how before you can determine what exactly, if anything, intel is deficient at. apparently, the previous version of geekbench included SHA1 as a large part of the integer testing. all you need to have a huge score in that is a small fixed-function unit. apple had it, intel didn't. also, seems like a lot of the tests were extremely cache friendly, so didn't test the greater memory subsystem at all.
On JavaScript benchmarks like octane and antutu html5, A11 clearly gave intel i5, and even i7s a run for the money. I tested it personally on my laptop which runs a skylake i5u
Exactly, much of the circuits in x86 cpu are there just there to maintain backward Compatibility. Even if no users today ever use it, they are still required to support OS and codes from eras like the windows 98ARM is much more suited to low power solutions and isn't dragging around a 30+ year old architecture, so it's going to have an edge in this niche. That Intel does as well as it actually does in those market segments is a testament to their engineers.
Also, please find me some posts from people who are claiming that SoC performance would stop. I seriously doubt these exist in any quantity or are coming from anyone that the regulars here treat as an expert.
Alright, if you wish to compare them like that, we can eliminate the software altogether from the equation and directly compare flops between the two processors. But again, we will have some people disatisfied and claim that flops can hardly translate into 100% real life perfomance(which is true even though hardwarely, more flops means indeed faster computation per seconds)Running a javascrupt benchmark on your iPhone is not directly comparable to running the same javascript benchmark on your laptop. The benchmark is running under two different operating systems that each have their own overhead. Windows overhead is MUCH HIGHER than iOS or Android.
If you try to run Windows 10 on the A11, it's not going to perform like an i5 and definitely not like a high end i7. Snapdragon 835 running Windows 10 is giving the same performance as a 6W celeron. http://www.pcgamer.com/leaked-benchmarks-of-windows-10-on-arm-are-not-impressive/
A11 is faster, so maybe it creeps up into low end i3 performance. Maybe. Point is, you can't easily compare the benchmark scores from one platform to another, especially from a very light weight phone operating system to a very robust desktop/enterprise operating system like Windows.
Performance per wat only improves at the sweet spot so it's only relevant to portable devices,it doesn't affect top(as in unlocked high clocked CPUs) performance.If 3nm is true, how much more performance per watt and number of cores can we expect with x86?
I was thinking of an 8-core CPU at 3600Mhz with a 65w TDP.Performance per wat only improves at the sweet spot so it's only relevant to portable devices,it doesn't affect top(as in unlocked high clocked CPUs) performance.
As for the second part of the question,as AMD just showed us the only limitation to how many cores a company can fit into a CPU is the amount of desperation they are in.
If you try to run Windows 10 on the A11, it's not going to perform like an i5 and definitely not like a high end i7. Snapdragon 835 running Windows 10 is giving the same performance as a 6W celeron. http://www.pcgamer.com/leaked-benchmarks-of-windows-10-on-arm-are-not-impressive/
There is no magic ISA. If it were possible to make a RISC CPU at 35W that performed the same as an 8700k at 91W, it would have been done
Geekbench scores:
A11 (6c/6t@ 2.4 GHz):
Single core: 4200
Multi-core: 10,000
i7-8650U @2.11 GHz (4c/8t)
Single-core: 3900
Multi-Core: 13,000
but that's not an ISA discussion.It's possible, if the 91W CPU is made by a team that's less capable than the team that makes the 35W one. Also, power scales drastically lower than performance, so just 10-15% sacrifice in performance may get from 91W to 35W.
Even if Moore's Law scaling completely dies, there's still the effect of good execution by level headed management and realistic goal setting.
with apple and its custom ARM cores i think we have to be especially careful trying to separate out what the general execution units in the cores are capable of vs. what is the benefit of vertical integration in a computer company. apple almost certainly carefully profiles iOS and its loads and builds the processor with that in mind (and vice versa). i wouldn't be shocked to find that apple's chips have more fixed function units in them than anyone else, in order to accelerate common tasks with a minimum of power consumption.On JavaScript benchmarks like octane and antutu html5, A11 clearly gave intel i5, and even i7s a run for the money. I tested it personally on my laptop which runs a skylake i5u
Aren't we pretty much there already with Zeppelin? I mean the 1700 is only off by 300MHz.I was thinking of an 8-core CPU at 3600Mhz with a 65w TDP.
Nope off by 600Mhz:Aren't we pretty much there already with Zeppelin? I mean the 1700 is only off by 300MHz.
Nope off by 600Mhz:
https://www.newegg.com/Product/Product.aspx?Item=N82E16819113428
I doubt TDP is calculated with all the cores at turbo clocks.All core turbo is 3.3GHz. So it runs all cores without modification at 3.3GHz at or under 65w assuming proper cooling.
I doubt TDP is calculated with all the cores at turbo clocks.
Intel says base clocks and AMD is very vague.
What are the clocks at 100% load for the 1700? I have seen the power figures somewhere, but I do not recall the site listing the clock speeds. Tom's shows 82 watts for the torture loop, but doesn't give the clocks. Reviewers don't seem to list the clocks at a given power consumption level for the desktop chips they review, or whether the clocks are sustained.The tests I have seen have 100% load on a 1700 at being at or below 60w. Intel's numbers are based on full AVX2 load which basically will only run at base clocks. So a 7700 for example running at all core turbo is probably under TDP.
TDP isn't power usage. It's a power design for cooling requirement. But on Desktops it's going to be pretty damn close to power usage. I know the 1800x can be on the other side of it though pushing 95-100w at 100% CPU usage.
The thermal design power (TDP), sometimes called thermal design point, is the maximum amount of heat generated by a computer chip or component (often the CPU or GPU) that the cooling system in a computer is designed to dissipate under any workload.
The TDP is typically not the largest amount of heat the CPU could ever generate (peak power), such as by running a power virus, but rather the maximum amount of heat that it would generate when running "real applications." This ensures the computer will be able to handle essentially all applications without exceeding its thermal envelope, or requiring a cooling system for the maximum theoretical power (which would cost more but in favor of extra headroom for processing power).
Since safety margins and the definition of what constitutes a real application vary among manufacturers, TDP values between different manufacturers cannot be accurately compared. For example, while a processor with a TDP of 100 W will almost certainly use more power at full load than a processor with a 10 W TDP from the same manufacturer, it may or may not use more power than a processor from a different manufacturer that has a 90 W TDP. Additionally, TDPs are often specified for families of processors, with the low-end models usually using significantly less power than those at the high end of the family.
The dynamic power consumed by a switching circuit is approximately proportional to the square of the voltage:[7]
where C is capacitance, f is frequency, and V is voltage.
What are the clocks at 100% load for the 1700? I have seen the power figures somewhere, but I do not recall the site listing the clock speeds. Tom's shows 82 watts for the torture loop, but doesn't give the clocks. Reviewers don't seem to list the clocks at a given power consumption level for the desktop chips they review, or whether the clocks are sustained.
Does the I5-8400 actually drop to base clocks with an AVX2 load, for example?
That seems to be a mobile chip thing.
TThermal design power
There is no set standard for TDP, and it depends entirely on what the manufacturer decides constitutes a XXW TDP workload.
Intels definition of TDP from it's product specification:
"Thermal Design Power (TDP) represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload."
Key points:
Average Power (not max power)
Dissipated at BASE FREQUENCY (not turbo frequency)
and the most relevant, INTEL DEFINED WORKLOAD
Note: high-complexity doesn't involve AVX workloads!
T
Yes, we have discussed the definitions of TDP used by AMD and Intel many times.
If Intel desktop chips ever run all cores at base frequency, I can't ever recall seeing it happen.
We don't know what loads AMD or Intel use to set the TDP.
but that's not an ISA discussion.
One contributing factor is that there's less competition. There are only two real x86 core designers: Intel & AMD. And only one with any real R&D expenditure. There are a boatload of ARM core designers: ARM, Apple, Cavium, Nvidia, Samsung, Qualcomm and more. And a lot of them are investing significant amounts of money into core designs.
Some caveats: prefer SPECint/fp than Geekbench. And I'm not seeing someone measuring the actual power draw of the A11 during those tests.