Aah okay. Thanks for answering-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
AMD THREADRIPPER VS INTEL SKYLAKE X
- Thread starter faceman123
- Start date
Aah okay. Thanks for answeringStop going over the same old ground over and over and over and.....
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,333
- Location
- Planet Earth
Nope - if you are that worried about I/O you would be using PCI-E based ones not SATA,so basically you are saying that the scenario you are talking about is an artificial one,where someone needs massive IO performance but CBA investing in PCI-E based cards.
This would indicate someone on a budget,so if they can't afford PCI-E drives that means it is unlikely they want to pay the price for a more expensive Intel setup.
Look at this part of an earlier post:
So an over 5 times speedup going from SATA to PCI-E.
If the platform "falls flat on it's face" in the operations you mention, then i'll take my platforms in future with my nose pressed firmly to the ground.
Would need the equivalent drive on Intel to compare against.
I don't think you are understanding the potential issue I'm talking about.
Last edited:
Him too eh? it seems that no one does.
One would have thought that if it was as serious an issue as you are saying it wouldn't need much explaining, it'd be obvious to anyone.
Last edited:
^^ I don't personally have a need for more than what I'm getting and I'm getting the full extent of my hardware - my story is irrelevant to the needs of an enterprise type environment or those doing other tasks where lots of cores and storage IO comes into play. Which is where Threadripper comes into the story.
Not sure if he has RAPID mode on or not but it can vastly inflate the numbers for short tests:
Holly #### IOPS on that ^^^^
Not sure if he has RAPID mode on or not but it can vastly inflate the numbers for short tests:
For a start these are two completely different platforms, the one you are complaining about is X370, a mainstream desktop platform, this is the Threadripper thread, which is an enthusiast platform based on a server platform, that is very dependant on IO throughput.
What you doing here is trying to exaggerate a quirk on one platform and then attribute to a completely different platform we know nothing about.
Also, when you use extreme or hyperbolic language like "falls flat on its face" it just makes your whole argument look like its shilling or hate agenda driven.
Your whole demeanor and persistence in here just smells extremely suspect.
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,333
- Location
- Planet Earth
Things like SATA Express felt flat on its face even though it was faster in every way and it had very limited market penetration it seems,and PCI-E based storage is not as common on HEDT based platforms as people on forums think. IO is important but if it were that important you wouldn't be using SATA3 which has long been considered a major bottleneck and SATA3.0 would have long disappeared off Intel HEDT motherboards.
That indicates for me that a decent amount of the HEDT market is either not as IO limited as we are led to believe or cost is a significant barrier. This is the whole problem - looking at where Intel is pitching its CPUs and the rumours of where AMD is going to start pricing 16C Threadripper it seems really weird if you are willing to pay significantly more for a 16C Intel CPU you would hobble it using a SATA3 based storage solution for IO intensive usage.
Also looking at some instances of workstations I know friends have used in research/commercial instances which are more IO limited,they all have PCI-E based systems AFAIK.
But it is only a certain percentage of systems - its like all these people talking about overclocking HEDT CPUs.
In all the large scale deployments of HEDT based desktops I have seen,all of them were running at stock clockspeeds.
Edit!!
I think that is another aspect we forget - I doubt the large scale purchasers of these systems,and whatever AMD does Intel will probably still sell more,will be running the top SKUs,fastest RAM possible or even the most expensive motherboards.
That is because most will be sold by companies like Dell,HP etc who are selling these in bulk and tend to be very conservative with hardware.
Last edited:
I'm getting the fullspeed out of my hardware and I'm not doing things where enterprise level of storage IO is required - not sure what is so hard to understand unless you are completely missing the point of what people like FredFlint is saying.
Someone using a similar drive as me and getting around half what their drive should on paper while I'm getting pretty much the expected level that is less than ideal even if it just means they are seeing a few seconds longer load times, etc. but if that does translate to the high end/enterprise level CPUs that becomes a much bigger deal - especially as HEDT type environments is where those broader areas of storage IO that are identified as not doing great on Ryzen in the articles I linked become more significant.
I'll say again I acknowledge that we have no idea how these areas will perform on Threadripper when it is released but it is still based off a platform we can see potentially less than ideal aspects in that do have relevance to that area of computing.
Not sure if he has RAPID mode on or not but it can vastly inflate the numbers for short tests:
I think that answers your question.
That is a shame for all of the 5% of scenarios where RAPID mode actually does anything tangible. I'd forgotten NVMe platforms don't support it.
You are getting more and more confused - where did I suggest (or not suggest) using SATA on a HEDT setup? and you might as well ask me why I'm not using an array of RAID0 NVMe drives - its irrelevant to my needs and irrelevant to the point I'm making - as you don't understand that its not surprising you don't understand the point I am making about storage IO performance.
That's because they don't need to. Could also explain why i don't feel the need to be flat on my face.
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,333
- Location
- Planet Earth
Because again I actually know people in realworld situations which are IO limited which use PCI-E based storage - lots of the workstations out there are still on SATA,many are not going to be on the latest drives,or even the latest chipsets either,ie,lower performance than something newer. These systems will only be replaced when the support contracts are up - if IO is so significant then they will all be specced with PCI-E based storage but many probably won't be.
Also the market does not agree with you - SATA Express and its drives were not a massive success and probably most HEDT based desktops are still sold with much slower SATA3 based drives with PCI-E ones as an option.
Plus you missed the relevant part of his post - an over 5 times speedup going to PCI-E based storage.
Then you also don't seem to understand that most workstations will be sold by large scale integrators like Dell,HP,etc who are very conservative in their hardware choices including RAM speed,storage speed,etc. So we might measurebate over what is faster,which is worth diddly squat if the integrator does not ship the systems 3GHZ RAM or the latest Samsung consumer drive.
The problem is we are on a enthusiast forum and - this is where everybody is running 3GHZ memory on their SKL/KL CPU overclocked to 5GHZ,but the reality is literally most people I know who have a Core i7 have it at stock.
it's not even ryzen users beta testing, there's a few things overall
the imc has a weaker design compared to intels, it's ridiculously easy to crack past 4000mhz on 4 sticks with Intel, but for some people can't even get past 2666mhz with two sticks even with the newest aegsa update.
so they really need to address this, which isn't possible with threadripper as the chips as essentially r7s stitched together.
half of the launch issues with ryzen were amd gave them no time with the cpus to make more matured bios', motherboard makers were apparently pleading with amd not to do this with threadripper at computex (according to gamernexus and Jayz when they spoke to vendors)
infinity fabric is a double edge sword, yes it does work well for stitching cpus together but it's inherent fault is being tied to ram speed, which may seek ok for the high end desktop overclocker, but the vast majority of people that don't overclock will be usong 2133mhz ddr4, whoch bottlenecks the infinity fabric much more, causing the cpus performance to call behind Intel's even more.
with zen 2 (not zen+ as that's just a refresh like the rx580) I would expect they need to have a look at infinity fabric, and try to find a way to increase its interconnect speed without relying on ram, if they could somehow (I have no idea how this would work) but have it work at say, the speed of 4000mhz ram that would be far more effective for the average user and any task that runs across multiple cores.
LOL back to this of inventing a massive amount of things you think I said - that I did not say at all. How is SATA express and its success or otherwise relevant to anything I've been saying? (go back read my posts there is nothing specific to SATA or PCI-e/NVMe I'm talking about the storage IO capabilities as a whole which appears, though its a mixed and far from straight forward story, to be showing in all types and areas) sure you get a speed up going to PCI-e based storage but the articles still show areas of not so great performance compared to the on paper specs for storage and compared to what Intel platforms are managing while as you said in the real world people are often IO limited on these platforms so they will want to get the best out of their hardware and so this could be a significant point for a CPU that is designed for those environments. (This is obviously an assumption that what is seen with Ryzen so far translates to Threadripper which remains to be seen).
Do you not think with this being the first generation of the Zen architecture that they will be looking to improve in areas that need improving on? I mean how long has it taken Intel to get to the point they are at now with the 7700k and 6900 ect? And as it stands Ryzen can already trade with them at various points with threadripper set to contest a largely uncontested market for Intel. Would you not say that AMD are in a much better position now than they were a year ago and they've yet to release their apu's of which Intel cannot match?