Posts by Fenton 271 publicly visible posts • joined 19 Feb 2008
Entry level Mac mini is a great machine for the price, however upgrades to future proof the machine are just stupid, £200 per 8 GB of Memory, £200 to go from 256GB to 512GB storage is just stupid and you can't upgrade it at a later point.
This is likely the biggest sticking point for people likely looking at the Mac. It's a pity the PC ecosystem does not have suplier with the same design flair.
Yes it looks like a bit of a bug in the OS, the real issue for AMD is pricing of the lower end parts.
At the high end, i.e. 7950X they still beat the 13900K in a lot of gaming and productivity benchmarks (a big fix will hopefully help them get over the CCD issue).
It's the 7700x and 7600x that need at least a $50 price drop.
However the price is worth it from a long term investment point of view. With AM5 we have at least two more CPUs we can drop into the motherboard. With 13th Gen intel a new motherboard will be required to go to 14th/15th gen.
From what I can gather from various rumors and Roadmaps 8th Gen Ryzen will have new Power cores and 4th gen Cores will be used as efficiency cores (albeit with less cache, i.e. Zen 4C).
This card is not a competitor for the 4090 but a competitor for the 4080 and 4080ti.
Rumor has it this card could clock to 3Ghz, some AIBs may bring out cards with 3x 8 pin power connectors.
The race will be interesting, given we all have to tighten our belts I think AMD are going for bang per buck rather
than best performance, but given the chiplett design they could quite easily come out with a faster chip
It looks like the 7000 chips are a bit of a monster and in ECO (65W) mode can get 80% of peak performance. And they have not yet announced the 3D cache versions. They also have room to drop prices, 13th Gen may well be a difficult sell especially as there is no upgrade path as 14th Gen will be a new socket. AMD will have 8000/9000 series on the same AM5 motherboard
As a persistent memory architecture it was just not fast enough and useless in large virtual clusters where you have no idea where the VM is likely to restart. So is was really just a bare metal solution only.
Memory hungry applications tend to be memory bandwidth and latency dependent, so to get the required performance you ended up having to run in a 1:1 DIMM to Optane configuration.
As a storage solution (sitting on the PCIE BUS) is was far too expensive and again was local to a physical machine, so at best it was a good caching mechanism before writing to a fast remote storage device, but then again, an array of fast nvme drives would have been just as fast and cheaper for the same capacity.
mmmmm AMD already have surpassed Intel on core count and power efficiency in the DC.
Milan vs Ice lake single core performance is roughly on par, with 64 Cores vs 40.
Sapphire Rapids is based on the Alderlake chips which on the desktop only just beats AMD in single threaded performance.
Current leaks of Zen 4 show a minimum (AMD sandbagging) 15% ST improvement and Genoa will be up to 128 cores per socket (96 pretty much confirmed).
Sapphire Rapids = 8 Memory Channels
Genoa = 12 Memory Channels
The only advantage I can see with intel will be in the 4/8 and beyond socket market (needs 4 sockets to compete with AMDs 2 socket)
The Epyc chips are a thing of wonder though.
The compute density is off the charts, even Sapphire Rapids won't compete here.
It looks like Zen4 based Epyc processors will have upto 96 Cores per socket, giving you 384 vCPU for a 2 Socket Pizza Box.
The only place where intel can compete is in the 4 Socket + area for very large memory configurations, even then it looks like a 2 socket Genoa could handle 12TB of memory assuming 1 slot per memory channel and 512GB Dimms.
Come on AMD produce a 4 Socket Box, that will even eat into the Intel 8 Socket business.
A standard single app is fine, say gimp or libra office.
The second you want to do anything out of the ordinary, say set up a development environment, it's one package after another.
I'm a noob programmer (Basic and Fortran from the 80s is where I cut my teeth).
On Windows I just install Visual Studio and have a single environment with everything I require, right there. The programing language, the Editor and any dependencies, I can then download any add-ons via a menu and the entire environment is set up for me.
On linux, setting up something as a simple Python development system, involved command line after command line of installs.
To make headway on the desktop, Linux needs to standardize on a single desktop environment that looks good out of the box on every distro.
Yes make it customizable for those who want to, but the majority of end users will at the most want to change the background picture and that is it.
It needs a standardized App store that is cross Linux compatible, i.e. double click the app you want to install and the store automatically determines
which distro you are using.
And fix application dependencies. I use various UNIX/Linux/Windows/Mac computers on a day to day basis.
90% of the pain on Linux is fixing apps at the command line to install dependencies as all the documentation, discussions all focus on fixing via
command line. The average user does not want to touch the command line.
This is always a business issue. Bob from accounts insisting 15 year old data has to be online and nearline is not good enough. Because he has to run that 15 year old badly written report, despite the fact the data already exists in that expensive Data warehouse and analytics tool.
Always in the too hard to break through the politics bracket.
The CPU count and performance is not actually interesting.
What is interesting is all of the other accelerators, which may mean more people move over
to the apple ecosystem. This may worry Microsoft.
Once we get some real world benchmarks and these machines prove to out perform more expensive
intel equivalents, they may well cause a shift to apple which will hit Microsoft/Intel/Nvidia/AMD.
Hopefully this will spur the traditional makers to start implementing accelerators/extensions into x86
machines.
It will be interesting to see where Apple go with the next gen PowerMac.
Are they really going to produce an SoC with 2TB of Memory, or will they go
with an expandable slot design and PCIE slots for dedicated GPUs?
And what about devices that sit between the PowerMac with its stupidly high prices
and the iMac pro/Mac Mini. I'd still love to see an expandable Apple desktop, that can grow
with me/my business.
if there is one think that annoys we with Windows 11 (although has been very stable whilst I've been testing), is the change in right click behaviour.
Since windows 3.1 Cut/copy/paste was always in the same same position on the right click menu.
Now there is either an icon at the top or the bottom of the menu or you have to go into the more options menu.
This has slowed me down quite a bit (yes I know there are keyboard short cuts).
As somebody who has used pretty much most version of Unix (Solaris,HP-UX,AIX,BSD,SCO, etc), OS-X and Windows since the 90s either in a professional or personal capacity, my conclusion is that Unix/Linux, is developed by developers for developers or highly experienced IT people. Yes from a functional point of view they are far superior being far more flexible, but there are also issues with that in itself.
Even an experienced user can make fundamental mistakes as their knowledge is based on experience. E.g. used to configuring a VM to run with a Netapp filer, suddenly presented with an EMC Isilon, will make the wrong assumptions and wonder why things run slowly.
But on the personal PC front, the end user should be at the forefront of the developers focus. They want to install the OS without any issues and the hardware should just work (which it does the majority of Windows installs, because of third party support).
They want to install apps without the need to go down to the cli (which is still to often a requirement) and run them.
Apple have managed to do this on top of OpenBSD, the Linux community needs to stop the UI wars and focus on a single UI that looks great and is easy to use for an end user.
By all means have all the useful stuff, but hide it away so it doesn't confuse the end user.
The only piece of Enterprise software I know that does not run on AMD is SAP HANA, but this is purely an artificial limitation imposed by SAP (I have heard AMD themselves are running HANA on Epyc).
To get this market and in light of Intels delays, AMD really need to think about the 4 Socket market for very large memory foot prints e.g. larger than 8TB, where they can then also start competing with IBM and Power.
AMD have always been great innovators, first to dual core, first to SIMD instructions, first to integrated memory controller, and first to x86-64, I see the as a modern day digital/Sun. Great innovators not neseccarily good at execution/marketing but Lisa sure is turning that around.
At least his idea of raking a forest was harmless.
His idea of buying Greenland was bonkers but again relatively harmless (diplomacy aside)
His idea of building a well, financially bonkers but not a killer (directly anyway)
But mentioning fish tank cleaner and bleach to cure Covid is beyond the pale especially as he knows
he has a lot of uneducated supporters (not showing your love of them here are you Trump)
Surly somebody in his family has to guts to tell him to think before he
opens his mouth.
Weight problem aside you can increase efficiency quite a bit by having a central engine charging a small amount of batteries which power electric motors. The problem with gas turbines is they are very efficient at constant load (e.g. cruise), but quite fuel hungry at full boost (take off/Landing) also they don't respond. that quickly to changes in speed.
Whereas an electric motor responds very quickly to speed changes and if you have enough battery power (generated during cruise or charged on the ground) for the high thrust parts of the flight envelope you can increase efficiency quite a bit.
Needs a new generation of solid state batteries to work properly in an aircraft setting mind.
There has to be room for something to sit between a properly developed app and a spreadsheet.
Myself as a non-developer have created a number of "applications" in excel, only to find out that a spreadsheet is totally the wrong tool.
It needs to be duplicated again and again for each project (surely just needs a project key field), then you have endless tables for lookups (mmmm maybe a relational database is needed), very quickly the spreadsheet goes out of control taking 30mins on a fast machine, because I've updated one key field and 1000s of rows and 10s of tabs will suddenly be updated.
The closest I've found is "access" and Googles App Maker, but again if you want to do anything remotely interesting or outside of the tutorials or templates you need to be in full developer mode and any documentation for the intermediate user is non existent.
I'm still waiting for a cross platform environment for simpletons like me.
The closest I've found has been visual basic for the desktop and appmaker.
But I often feel developers develop languages for full time developers, Rather than somebody who just wants to build a quick and dirty tool as they have exhausted what they can do in a spreadsheet.
This just makes the situation worse.
A) Any goods coming from Mexico will just be more expensive for the average US consumer causing inflation which in turn will mean higher interest rates.
B) Hardship in Mexico will be even worse driving more people towards the boarder and the so called promised land.
Cisco do a C880 (which is actually a OEM Fujitsu), HPE (who now own SGI) and also Atos Bull all do 8 Socket boxes as well as IBM Power
But if you need any sort of high availability you will always need at least two.
If your workload can be split up much better going for smaller nodes with one spare.
These days with such high core counts it's generally the memory costs that determine the server cost.
Do you run a single server with high core counts and high memory requiring 64GB or 128GB Dimms or lost of small ARM servers with 16/32GB Dimms. All of a sudden real estate becomes an issue.
Having a common kernel is a good idea, what Apple should do is think about different personae depending on form factor which will have to be adhered to my software vendors, I.e. a cut down interface with different functionality in mobile mode and a full desktop experience when linked to a monitor and keyboard/mouse.
I actually quite like some of the touch features in windows 10 in desktop mode
Now Hyperconverged is something entirely different.
But converged/Hyperconverged also means, get all your support and lifecycle management from one place. Rather than constant finger pointing and time wasted working out which patches/firmware updates work together across different vendors.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
