We're likely to see 3.9-4.0 on 14 / 18 core configs.This is exactly what I feared... Does that mean the actual cap is lower than (or hopefully at least equal to?) 4.2 for the 14 and 18 cores as well?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
iMac Pro vega 64 or 56
- Thread starter benoitc
- Start date
- Sort by reaction score
Given battery-gate and everything going on with throttling iOS processors over time (with no advertised speed, mind you), does this qualify as false advertising if the default configuration fails to meet the claimed specs in ANY scenario?
If it does, it's on Intel's side. Apple's not responsible for parroting what Intel tells them the CPU does.
This may be true, though if the processors are already down-clocked in the base frequency, you might think they would know the conditions of operation when installed, and whether or not they may affect the product claims.
In other words, it's fair to say "apple isn't responsible for advertising the specs of a supplier's product" but isn't it also fair to say "apple is responsible for acknowledging limitations of a component in their allowed usage?"
I would expect if you used software to disable 9 cores and run a single core at 4.5ghz, apple would say you violated warranty, which is probably reasonable of them to claim.
However, claiming a processor is capable of speeds it is specifically unable to hit, even in an isolated un-realistic test... that sounds a lot worse than slowing down a processor that was never promised to have a minimum speed to me?
Not arguing with you, SecuritySteve, you are quite well informed and I appreciate your willingness to share the information
I see what you're saying. You're right that Apple definitely has it's hand in this cookie jar. I'm just not sure how culpable they are. I also didn't read into any tone you had, but I appreciate the comment.This may be true, though if the processors are already down-clocked in the base frequency, you might think they would know the conditions of operation when installed, and whether or not they may affect the product claims.
In other words, it's fair to say "apple isn't responsible for advertising the specs of a supplier's product" but isn't it also fair to say "apple is responsible for acknowledging limitations of a component in their allowed usage?"
I would expect if you used software to disable 9 cores and run a single core at 4.5ghz, apple would say you violated warranty, which is probably reasonable of them to claim.
However, claiming a processor is capable of speeds it is specifically unable to hit, even in an isolated un-realistic test... that sounds a lot worse than slowing down a processor that was never promised to have a minimum speed to me?
Not arguing with you, SecuritySteve, you are quite well informed and I appreciate your willingness to share the information
This reminds me of a similar case that was brought against Apple a few years back. There was some guy arguing that because the OS took up space on an iPhone with X capacity of a hard drive, Apple shouldn't advertise that the capacity was X, but instead X - OS due to the real world limitations of the drive. I'm not sure I bought it even then, but it is an interesting parallel.
I remember that! If I recall, that minor grievance exposed a larger issue, that apples "GB" was actually 1000 mb, rather than 1024, further reducing capacity from the expectation.I see what you're saying. You're right that Apple definitely has it's hand in this cookie jar. I'm just not sure how culpable they are. I also didn't read into any tone you had, but I appreciate the comment.
This reminds me of a similar case that was brought against Apple a few years back. There was some guy arguing that because the OS took up space on an iPhone with X capacity of a hard drive, Apple shouldn't advertise that the capacity was X, but instead X - OS due to the real world limitations of the drive. I'm not sure I bought it even then, but it is an interesting parallel.
I think having the OS installed reducing advertised capacity is a bit of a stretch, you'd lose that capacity anyway, no matter which disk you install, or which OS you install on it...
However, the fact that an apple 200gb disk had less than a Samsung 200gb disk, both when formatted and unformatted... that struck me as a bit more shady, even if the difference was smaller than the installed OS footprint.
If the result is apple adding a disclaimer for every "SSD Capacity" claim with a footnote "1GB = 1000MB, not 1024 MB" (which I have actually seen), I think it should be reasonable for apple to say "1ghz = speed measured before apple down clocking" or more realistically "actual speeds may represent a 5% loss due to thermal constraints, disabling cores in system preferences can achieve advertised speeds"
But we know that as long as they keep selling computers, they can say whatever their lawyers tell them they can get away with.
I wonder what would happen if anyone tried putting a regular Xeon W in the iMac pro socket... if it would naturally down clock in software, fail to boot due to security or compatibility, or just run hotter?
P.S. SecuritySteve: I just saw in your sig that you have an 18 core iMac pro... is that on order, or do you have an early unit by any chance? You don't have to answer if it might get you in trouble
Ah yes, I remember the part about the whole "GB" issue. I seem to recall that it had something to do with the format table not being included in Samsung's disk capacity advertisement ... but my memory is being stretched to it's limit now!
From what I've heard, the regular Xeon W would work, it just would run hotter and cause the machine to be more noisy. There may also be some slight system instability, depending on the chip. These are typical problems that you also encounter on Hackintosh builds.
Regarding Bootcamp: Does MSI Afterburner (overclocking, undervolting) or Mac Fan Control work? I saw an iMac Pro reviewer on youtube claim that the Vega Pro cards in iMac Pro are "locked in", and you can't tamper with undervolting and overclocking, or change the fan speed. Is this true?
I've read AMD using unnecessary amounts of power to run the Vega GPU, and if undervolting is an option - the Vega 56 or 64 could be tweaked for higher performance at the same thermals.
I've read AMD using unnecessary amounts of power to run the Vega GPU, and if undervolting is an option - the Vega 56 or 64 could be tweaked for higher performance at the same thermals.
Last edited:
Intel doesn’t publish all-core turbo specs anymore, just base frequency and single-core.
I’m unsure what the base frequency spec is good for, besides establishing the TDP. It’s not a particularly relevant spec for end users, since with typical cooling the base clock is well exceeded.
Additionally, the single-core speed is more an Intel marketing gimmick so they can advertise the highest possible clock speed attainable (for a completely unrealistic and almost-never-seen potential workload). So also not very useful to end users.
I get that CPU performance is workload-dependent, but the two specs published by Intel are just not that meaningful.
Irrefutable logic "You chose WISELY"
Also note, look at the cost of the Vega 64 as a GPU card for the PC.
On amazon the cost of Vega 64 is easily 2 to 3X the cost of what you would pay apple to upgrade.
So if you add it later as a external eGPU option you really are better off buying it as an internal option.
The true handicap of this $5000 USD system is not the number of CPU's but it's the GPU.
I have added the V64 to my purchase.
This is becoming more and more true as prices rise for these cards. Miners are driving the GPU market insane, I wonder if it's a bubble that will impact vendors of GPUs when crypto mining starts to bust.
Thanks, appreciate it.
You wouldn't have the memory clocks too? Vega 64 is constrained by memory bandwidth, so would be useful to know the memory speed - regarding performance difference between the Pro 56 and Pro 64...