If implemented properly by the WiFi Alliance, major numbers should not change, but expect a barrage of :
WiFi 6.3++ultra/hyper/mega/advanced type branding on the APs and Cards.
Just like with 3G, 3.G, 3.5G+, 4G, 4G Advanced, etc...
All kidding aside, I think this is a good development.
Leave the IEEE 802.11letter(letter) to people that understand it, and give joe consumer an easier thing to go by, especially since WiFi was (and still is) a marketing name anyway.
Actually, I can not understand what took the WiFi Alliance so long...
On a more serious note, Good to know is working without major snafus.
But it will be a pain in the ass for me the next time I boot the bootcamp partition on my main laptop, or the old emergency/media laptop that is booted once a week.
I'll try darned hard to schedule the update in a convenient timeframe, if projects permit.
* I guess it means Professional Field Engineer, but the BOfH reference was ripe for taking...
When I set on my Mac and Fire up Powerpoint to make a presentation on how to install fusionsphere/Openstack on Huawei servers all while reading the manuals in fucking Hedex (windows only), while preparing hadout notes in word..., or when I sit in my mac and again fire up Powerpoint to preprare a rpesentation about how to instal, configure and administer the Hadoop backend for Nokia's CEMoD 16 while reading the manuals in AdobePDF (thank god), and write handouts in word, I am creating content too.
I guess the cute girl in Admin creates content when she fills up an excel sheet with my perdiem, flight tickets and hotel reservation to send to logistics to arrange everything....
Not only hollywood creates content, and not all content is ausio/visual.
Well, the free upgrade offer ended on July 29, 2016. At that time, Windows 10 had around 350 Million Devices.
Let's assume every single one of those devices were a non-voluntary unsuspected user sneaky upgrade.
that means that they had from then until now, 350Million more devices.
And no, you can not substract business or governments, either that do not give employees a choice, or that they do not have a choice themselves, lest you want to substract places like governments in Venezuela, Italy, brazil, and india, or many a company, where linux is forced, and you can not choose either.
At the end, the owner of the business choose voluntarily the tools to use.
It would be like if a police officer complained that the department uses ford cars instead of Chrysler, or if in a mechanic shop the mechaincs complained that the tools are craftman instead of stanley... (to honor the long standing slashdot tradition of automotive analogies)
In the end, OSs are tools, each with strengts and weaknesess, and should be used as such.
OSs are not religions or sports teams...
PS: Mac on my desktop, android on my pocket, Linux in my datacenter, if I can.
Yes, 700 Million devices, a good % of those devices are capable of Content CREATION/PRODUCTION, unlike most of those Android/Linux devices which are only capable of Content COMSUMPTION.
Having said that, I use a Mac on my desktop, and an Android on my pocket. And I advocated Linux in the telco datacenter A LOT (in late 90's early 00's time frame, when it was quite harder to argue the case against the Likes of Solaris, HP-UX and AIX)...
But yes, to each his own, this milestone is a good one for Win10.
OSs are like tools. Each has its streghts, weaknesess, and purpose, so should be used acordingly. Instead of treating OSs as religions, or sports teams...
Transistors are doubling every 24 months or so, on par with moore's original enunciation of the law, and slightly off the 18 months of his revision of said law.
What is not working anymore is _*"People's Interpretation"*_ of said law, that dictate that computers sould be 2x faster every 18 months. Moore never said that. He only said that in a given sqr centimeter of silicon, the optimum number of transistors would double every 24 months. then he latter revised the number to every 18 months.
When Moore's law was in full swing, say, in the 80's and 90's, making transistors smaller every 18 months allowed computer and processor architects to make the transistors switch faster, consume less power, we could pack more features, and make the processors cheaper, all of that at the same time. So, we had a Mhz race, on top of wich we integrated things in the processor, first paging unit, the Caches (L1 and latter L2) and Math coprocessor, then the memory controller, and more pipelines (starting with the UV pipes) and we depeloped new features/functionality, like protected mode, IA32, MMX, SSE, AMD64, et al, all of this at the same time. Please notice that all of that was made possible by moore's law, because, to implement most of those things, one needs, you know, more transistors...
At some point in the mid 00's, we hit the first roadblock, and it was not possible to make the transistors switch faster without consuming a significant amount of power, so, one dimension less.
But moore's law did not stop. We still were able to double the number of transsistors every 18 months. That's why we were still able to pack more features (SSE*, AVX*, transactional memory, virtualization support, advanced Ecryption support) and more cheaper things (higher core counts, a GPU, PCIe controllers, ETH controllers and more) in the processor for the same price. But not with increased Ghz. And the SW world has been slow to make use of many of those new features/things, it seems that performance has stagnated. If moore's law had stoped, we wouldn't have been able to increase core counts (and cache, and all the other stuff) as we did, because all those things require, you know, transistors to implement...
Now, in 2018, finally, the REAL Moore's law is strugling and limping (ask intel 10nm and GloFo 7nm people). It has a few cycles left (ask the TSMC and Samsung 7nm people), but not many more.
In the meantime, yes, new architectures are worthwhile, fine and dandy (and more publishing and headline worthy), but a more inmediate thing would be to make better use of what we already have (SSE, AVX, GPGPU, Multiple cores)...
All of us played with margins, spacing, and fonts (to the extent possible) to make an essay look bigger, from the times of Typewriter, even going so far as to chose the typewriter to use among the three in my house to suit my needs, the most uncofortable one (but with bigger type) for essays with a set minimum # of pages, or the most confortable one for longer essays, or when there was no preset limit.
That's why, with the advent of computers, smart teachers request the work as a PDF and count words, not pages. Yes, a word count is also open to abuse, but less than # of pages alone.
Myself, I put a minimum AND maximum limit, both on pages AND on words.
and move their timezone to GMT. they currently sit on gmt + 2, which does not suit thir geography (it was done on the whim of a dictator to "show support for germany").
That, coupled with their weird times (3 hour lunch break starting at two) wreacks havoc on bussiness with the rest of europe, and the world...
I know a few people that were put off of apple's ecosystem because of a lack of dual-SIM phones.
There are a few reasons to have it, like people who hunt for the best price/coverage between two operators, or executves with work and personal numbers, or people who travel a lot.
whatever the reason was, if you wanted/needed the feature, now you have an excuse to go for an iPhone.
I love that someone with a UID as low as yours would actually include *BSD in a list called "Big Three".
How Slashdot has fallen...
AnonC, I slightly altered your Quote.
One one realizes that MacOS and iOS are kinda sorta BSD-ish behind the scenes, and that Android uses the Linux Kernel behind the scenes, one can be sure that those are indeed "The Big Three"
During my IMBA at Instituto de Empresa, we had a business case about Global brands which selected names that were downright offensive in other languages. Pajero was but one example of bad names in spanish. And there were examples in other languages too.
I guess that in a couple of years, Pocophone will be added to the infamous list.
I worked with a lot of Finish guys (Nokia, technomen) at the start of my career, even went to finland for a job interview in Jan 2001, but alas, it was not meant to be.
What you say abot performance/watt in ultra large scale deployments (Like Amazon, Azure, facebook, Google) is 100% true, and well known, but my deployments are more large scale (Telcos). And also, if I took it to that level, the point would be lost to people who are still thinkig at the level of a gaming desktop, or a tower server for a 5 people company...
As per the security, AMD (and all the others) has similar issues with speculative execution and side channel attacks, some have more issues, some have less issues, is just that Intel gets the bulk of the publicity for the time being, due to their large size.
I already told A LOT about me to Facebook, on my own volition. What they are able to collect with Onavo is peanuts in comparison.
I also tend to recomend Onavo, but ONLY in facebook. After all, if some contact of mine reads my recomendation in Facebook, it means that they too have given FB enough info so far that what FB collects through Onavo is peanuts.
I use Onavo in my android Phone. Onavo is easy to set up, easy to use, free, mantained by a well known company that will be here tomorrow, and let's me access sites my oppresive government deems unappropiate, as well as sites that are collateral damage of the censorship. I do not need to define the country I use to connect, nor am I torrenting, or streaming on my phone.
What's not to like? Spying? Again, I told FB all they need to know, whatever they gather through Onavo is peanuts.
On my Mac, on the other hand, I use ProtonVPN, where (I think) they are not spying on me, I can chose the country, I can Torrent* or stream if so I please, and have a VM with TAILS (yes, I know, TAILS on a VM is not the most secure, but I have no money for 2 PCs, and booting TAILS from a pen-drive is a pain in the nuts).
Moral of the story is: use the tool best fitted for the job.
And in the end, I'd rather have mark suckerberg spying on me than Nicolas Maduro...
* Right now, among other things, I am seeding Tails i386 2.12 (the last one with I2P), Latest tails, latest Kali, CrunchbangPP 32 and 64, and libreoffice win32 and LO for OSX...
"I favour AMD for servers (since this article is about servers) but there are many caveats beyond AMD's control that make it an uphill battle to use their chips."
the word "servers" is written twice.
The article is about a server chip.
I work with blade servers in telco. Also 4U ones with lots of 2.5" SSDs. Each of those servers cost $20,000,oo onwards.
If you want to discuss $200 mobos for gaming, is OK. I have been using and recomending AMD chips since the 486, K6-3, and now XEN. Skiped their K5, K6 and anythig buldozer derived. But that was/is for personal use. For server use, is a different story.
In my answer to fintux is my list of reasons. If you whish, go and read it. If not, no biggie, live a happy life.
I am preparing a blog post (+ LinkedIn series) about it. But since this is slashdot, and not something I write for CIO level people, and since you asked nicely (and are not an Anon Coward), I'll give you the TL;DR preview, I'll simply list the caveats, without much explaing. Please remember we are talking about SERVERS. For companies. Not desktops, not laptops, not gaming rigs, not servers for a home lab. Some people seem to forget that...
0.) If your workload needs anything Intel propiertary, like transactional memory, or some other weird stuff, then game over. By the same token, if you need something AMD exclusive, like the niffty feature that encrypts a VM's RAM so that not even the VMM can see it, then go AMD.
1.) Does your Workload (usually legacy) depend on single threaded performance? In that case, is likely that intel will be faster, but do a side by side trial and benchmark, as that will depend on the performance loss for the Spectre patches for both intel and AMD.
3.) Is your software licensed by core, not socket? (If it is FOSS, then the answer is, no, move on) Then, you need the configuration which achives the desired performance with the MINIMUM numer of cores, even if the Processor that achieves that is supper-expensive... The Software tends to be ORDERS OF MAGNITUDE MORE EXPENSIVE then the SERVERS (I am looking at you Oracle). So, again benchmark, but remember AMD has more cores, per socket, but intel tends to have more single thread performance, so, again, side by side trial and benchmark.
4.) Does your workload has problems with NUMA (also, mostly legacy)? In that case, certainly, intel. (with the use of "core-complexes" XEN has more "NUMA-ness" than intel, by a long shot. Even a single socket XEN processor is NUMA).
5.) If you are doing virtualization or cloud, does your VMM/Cloud SW support AMD properly? If not, is it worth it for your organization to add a new VMM to your stable on top of a new server chip? Like it or not, not all VMMs/Cloud SW support AMD yet, I hope this changes in the near future.
6.) If you need to LIVE migrate VMs from the old kit to the new kit and back, then the kit has to be either all intel, or all AMD. (migrating turned off machines is easier, the problem is LIVE migrating)
7.) Are the server makers in your approved provider list selling AMD servers? Is it worth it for you as a tech to go through the process with finance/purchases to add a new server provider?
8.) Are the server makers that sell AMD servers providing good service in your country? Are they providing good service to your company? Service varies by region and account.
9.) Are the server providers that deal with you providing AMD servers in the Form-Factor your project demands. I've seen mostly blades thus far. If you need 4U with many 2.5"disks, or 1U, or some other format, is it made by the server sellers your company approves?
Finally, a server is more than the microprocessor it uses, and actually, the microprocessor is a small part of the cost. Maybe the server seller is trying to get this project for strategic reasons, or increase their footprint in your company, or hit the sales goal so that they collect the bonus and pay their yatch. In the end, your project needs servers that cover it's needs in terms of security, performance, form factor and support, amog other things. If after all these hurdles are cleared, for whatever reason, a server seller is offering you a less expensive intel server that covers your needs, go for it. If the less expensive one is AMD, go for it
As you see, none of this is in AMD's control, it will get better over time, and none of this is due to evil machinations from intel.
I hope this is usseful to you, and hope I've got a new "friend" on/. Let me know what you think.
PS: For Grammar nazis, I bet my english is better than your spanish, AAAAND I also can communicate in french*, not too shabby for an electronics engineer.
*used to be able to speak, read and write french, but after 15 years of no practice, I forgot most of it.
If Intel completely reset the processor to its previous state after a failed speculative execution, then the processor would be immune to all speculative execution attacks--even speculative execution attacks that haven't been discovered yet. But it doesn't appear that that's what Intel is doing. Instead, their strategy appears to be to design a separate patch for each different speculative execution attack after that attack is discovered.
What did you expect? It takes about four years to engineer an new microprocessor architecture. More if you are starting from scratch. I am 99% certain that right now, there is a team of Intel Microarchitects designing such microprocessor from scratch, we may as well see the fruits of their labour in something like six years (maybe more, if there are more slowdownd in the Fab side of things). But in the meantime, Intel needs something to sell, otherwise, the company would go bankrupt before that new microprocessor design is out the door, so, we end up with this.
And before you think about going to some other manufacturer (AMD, Arm, IBM Power, Spark, VIA), remember that their processors are also suceptible to Spectre type attacks, so do a cost/benefit business case analysis before taking the plunge...
PS: I favour AMD for servers (since this article is about servers) but there are many caveats beyond AMD's control that make it an uphill battle to use their chips. This has nothing to do with obscure machinations from Intel like in the early '00s, and fortunately will get better in the future, but, again, the thing is, do a detailed cost/benefit analysis before taking the plunge.
In this decreases lifespan, but mantains, or actually increases, the number of healthy years in that decreased lifespan, is actually a welcome development.
No, is not made of tachions. The way standards work, is that drafts are circulated, and barring any major weirdass cornercases, what is ratified is pretty close, if not exactly equal tot he draft.
Remember all those "Pre-n" wifi APs at the end of last decade? Similar thing.
Therefore support for TLS 1.3 (as described in the draft) was added to firefox 52, and the draft was ratified into official standard in Aug 2018. No tachions involved.
If you mean unlocked BOOTLOADER, as in, I can install custom ROMs, than Sony is the one to beat. Just be careful of the specific model you select.
If by uunlocked you mean "I can stick a SIMcard from any opearator and the thing will work", then a Nokia 6 or higher is the best Android for that job.
Onorable mention for the MotoZ with Shatershield + MotoMods.
As I stated on my original post, I use Firefox ESR 60 on a mac. And Firefox on my android (KeyOne).
At home I use 9.9.9.9, 8.8.8.8 and 208.67.222.222 since I have better things to do than to set up my Synology to be my DNS server.
But when I travel, I use public wifi whenever I can get it, be that my hotel, the training centers were I teach, or, god forbid, a hipster coffee shop. And many of those need a captive portal to autenticate to the Wifi, and that depends on using the Network's DNS servers. So, I configured an "Automatic" setting on the network locales of my mac to handle those cases.
So, as a user of Firefox, I am not happy with this. I am capable enough to configure my DNS settings (or, if push comes to shove, set up a DNS from scratch, not even touching my nas).
So thank you for the inconvenience mozilla. I hope you guys enjoy the backslash when hipsters start to realize that they can not connect to the net in their favourite hipster watering spot because they can not get to the captive portals...
At least, the guys who use Mozilla in corporate networks will get this assinine setting turned off in group policies... as for the rest of us, a quick google and a trip to about::settings shal suffice
Slashdot Mirror
If implemented properly by the WiFi Alliance, major numbers should not change, but expect a barrage of :
WiFi 6.3++ultra/hyper/mega/advanced type branding on the APs and Cards.
Just like with 3G, 3.G, 3.5G+, 4G, 4G Advanced, etc...
All kidding aside, I think this is a good development.
Leave the IEEE 802.11letter(letter) to people that understand it, and give joe consumer an easier thing to go by, especially since WiFi was (and still is) a marketing name anyway.
Actually, I can not understand what took the WiFi Alliance so long...
PFE? Pinple Faced Engineer?
See "Bastard operator from hell"*
On a more serious note, Good to know is working without major snafus.
But it will be a pain in the ass for me the next time I boot the bootcamp partition on my main laptop, or the old emergency/media laptop that is booted once a week.
I'll try darned hard to schedule the update in a convenient timeframe, if projects permit.
* I guess it means Professional Field Engineer, but the BOfH reference was ripe for taking...
When I set on my Mac and Fire up Powerpoint to make a presentation on how to install fusionsphere/Openstack on Huawei servers all while reading the manuals in fucking Hedex (windows only), while preparing hadout notes in word..., or when I sit in my mac and again fire up Powerpoint to preprare a rpesentation about how to instal, configure and administer the Hadoop backend for Nokia's CEMoD 16 while reading the manuals in AdobePDF (thank god), and write handouts in word, I am creating content too.
I guess the cute girl in Admin creates content when she fills up an excel sheet with my perdiem, flight tickets and hotel reservation to send to logistics to arrange everything....
Not only hollywood creates content, and not all content is ausio/visual.
Well played Kaoshin, well played
Well, the free upgrade offer ended on July 29, 2016. At that time, Windows 10 had around 350 Million Devices.
Let's assume every single one of those devices were a non-voluntary unsuspected user sneaky upgrade.
that means that they had from then until now, 350Million more devices.
And no, you can not substract business or governments, either that do not give employees a choice, or that they do not have a choice themselves, lest you want to substract places like governments in Venezuela, Italy, brazil, and india, or many a company, where linux is forced, and you can not choose either.
At the end, the owner of the business choose voluntarily the tools to use.
It would be like if a police officer complained that the department uses ford cars instead of Chrysler, or if in a mechanic shop the mechaincs complained that the tools are craftman instead of stanley... (to honor the long standing slashdot tradition of automotive analogies)
In the end, OSs are tools, each with strengts and weaknesess, and should be used as such.
OSs are not religions or sports teams...
PS: Mac on my desktop, android on my pocket, Linux in my datacenter, if I can.
Yes, 700 Million devices, a good % of those devices are capable of Content CREATION/PRODUCTION, unlike most of those Android/Linux devices which are only capable of Content COMSUMPTION.
Having said that, I use a Mac on my desktop, and an Android on my pocket. And I advocated Linux in the telco datacenter A LOT (in late 90's early 00's time frame, when it was quite harder to argue the case against the Likes of Solaris, HP-UX and AIX)...
But yes, to each his own, this milestone is a good one for Win10.
OSs are like tools. Each has its streghts, weaknesess, and purpose, so should be used acordingly. Instead of treating OSs as religions, or sports teams...
Transistors are doubling every 24 months or so, on par with moore's original enunciation of the law, and slightly off the 18 months of his revision of said law.
What is not working anymore is _*"People's Interpretation"*_ of said law, that dictate that computers sould be 2x faster every 18 months. Moore never said that. He only said that in a given sqr centimeter of silicon, the optimum number of transistors would double every 24 months. then he latter revised the number to every 18 months.
When Moore's law was in full swing, say, in the 80's and 90's, making transistors smaller every 18 months allowed computer and processor architects to make the transistors switch faster, consume less power, we could pack more features, and make the processors cheaper, all of that at the same time. So, we had a Mhz race, on top of wich we integrated things in the processor, first paging unit, the Caches (L1 and latter L2) and Math coprocessor, then the memory controller, and more pipelines (starting with the UV pipes) and we depeloped new features/functionality, like protected mode, IA32, MMX, SSE, AMD64, et al, all of this at the same time. Please notice that all of that was made possible by moore's law, because, to implement most of those things, one needs, you know, more transistors...
At some point in the mid 00's, we hit the first roadblock, and it was not possible to make the transistors switch faster without consuming a significant amount of power, so, one dimension less.
But moore's law did not stop. We still were able to double the number of transsistors every 18 months. That's why we were still able to pack more features (SSE*, AVX*, transactional memory, virtualization support, advanced Ecryption support) and more cheaper things (higher core counts, a GPU, PCIe controllers, ETH controllers and more) in the processor for the same price. But not with increased Ghz. And the SW world has been slow to make use of many of those new features/things, it seems that performance has stagnated. If moore's law had stoped, we wouldn't have been able to increase core counts (and cache, and all the other stuff) as we did, because all those things require, you know, transistors to implement...
Now, in 2018, finally, the REAL Moore's law is strugling and limping (ask intel 10nm and GloFo 7nm people). It has a few cycles left (ask the TSMC and Samsung 7nm people), but not many more.
In the meantime, yes, new architectures are worthwhile, fine and dandy (and more publishing and headline worthy), but a more inmediate thing would be to make better use of what we already have (SSE, AVX, GPGPU, Multiple cores)...
JM2C, YMMV
... as we say in spanish speaking countries.
All of us played with margins, spacing, and fonts (to the extent possible) to make an essay look bigger, from the times of Typewriter, even going so far as to chose the typewriter to use among the three in my house to suit my needs, the most uncofortable one (but with bigger type) for essays with a set minimum # of pages, or the most confortable one for longer essays, or when there was no preset limit.
That's why, with the advent of computers, smart teachers request the work as a PDF and count words, not pages. Yes, a word count is also open to abuse, but less than # of pages alone.
Myself, I put a minimum AND maximum limit, both on pages AND on words.
and move their timezone to GMT. they currently sit on gmt + 2, which does not suit thir geography (it was done on the whim of a dictator to "show support for germany").
That, coupled with their weird times (3 hour lunch break starting at two) wreacks havoc on bussiness with the rest of europe, and the world...
I know a few people that were put off of apple's ecosystem because of a lack of dual-SIM phones.
There are a few reasons to have it, like people who hunt for the best price/coverage between two operators, or executves with work and personal numbers, or people who travel a lot.
whatever the reason was, if you wanted/needed the feature, now you have an excuse to go for an iPhone.
I love that someone with a UID as low as yours would actually include *BSD in a list called "Big Three".
How Slashdot has fallen...
AnonC, I slightly altered your Quote.
One one realizes that MacOS and iOS are kinda sorta BSD-ish behind the scenes, and that Android uses the Linux Kernel behind the scenes, one can be sure that those are indeed "The Big Three"
and other notable "influencers", because AI will be too busy fighting the "Grey Goo" predicted by Bill Joy and other notable "Influences"
That way, you can still be "whateverurweirdonameyouchoose"@"whateverservicewasavailable" and still be a very professional givenname.lastname@ieeeorg
Just remeber to make a google (or some other address) to match, since the ieee email is a redirector.
this will not impress anyone for or against you, but at least transmits a modicum of "seriousnes".
If you want to transtmit your funny side, that's what the personal interests part of the CV is for, mine lists scuba diving, cycling and reading.
During my IMBA at Instituto de Empresa, we had a business case about Global brands which selected names that were downright offensive in other languages. Pajero was but one example of bad names in spanish. And there were examples in other languages too.
I guess that in a couple of years, Pocophone will be added to the infamous list.
I worked with a lot of Finish guys (Nokia, technomen) at the start of my career, even went to finland for a job interview in Jan 2001, but alas, it was not meant to be.
What you say abot performance/watt in ultra large scale deployments (Like Amazon, Azure, facebook, Google) is 100% true, and well known, but my deployments are more large scale (Telcos). And also, if I took it to that level, the point would be lost to people who are still thinkig at the level of a gaming desktop, or a tower server for a 5 people company...
As per the security, AMD (and all the others) has similar issues with speculative execution and side channel attacks, some have more issues, some have less issues, is just that Intel gets the bulk of the publicity for the time being, due to their large size.
Cheers
I already told A LOT about me to Facebook, on my own volition. What they are able to collect with Onavo is peanuts in comparison.
I also tend to recomend Onavo, but ONLY in facebook. After all, if some contact of mine reads my recomendation in Facebook, it means that they too have given FB enough info so far that what FB collects through Onavo is peanuts.
I use Onavo in my android Phone. Onavo is easy to set up, easy to use, free, mantained by a well known company that will be here tomorrow, and let's me access sites my oppresive government deems unappropiate, as well as sites that are collateral damage of the censorship. I do not need to define the country I use to connect, nor am I torrenting, or streaming on my phone.
What's not to like? Spying? Again, I told FB all they need to know, whatever they gather through Onavo is peanuts.
On my Mac, on the other hand, I use ProtonVPN, where (I think) they are not spying on me, I can chose the country, I can Torrent* or stream if so I please, and have a VM with TAILS (yes, I know, TAILS on a VM is not the most secure, but I have no money for 2 PCs, and booting TAILS from a pen-drive is a pain in the nuts).
Moral of the story is: use the tool best fitted for the job.
And in the end, I'd rather have mark suckerberg spying on me than Nicolas Maduro...
* Right now, among other things, I am seeding Tails i386 2.12 (the last one with I2P), Latest tails, latest Kali, CrunchbangPP 32 and 64, and libreoffice win32 and LO for OSX...
In my original post I wrote:
"I favour AMD for servers (since this article is about servers) but there are many caveats beyond AMD's control that make it an uphill battle to use their chips."
the word "servers" is written twice.
The article is about a server chip.
I work with blade servers in telco. Also 4U ones with lots of 2.5" SSDs. Each of those servers cost $20,000,oo onwards.
If you want to discuss $200 mobos for gaming, is OK. I have been using and recomending AMD chips since the 486, K6-3, and now XEN. Skiped their K5, K6 and anythig buldozer derived. But that was/is for personal use. For server use, is a different story.
In my answer to fintux is my list of reasons. If you whish, go and read it. If not, no biggie, live a happy life.
I am preparing a blog post (+ LinkedIn series) about it. But since this is slashdot, and not something I write for CIO level people, and since you asked nicely (and are not an Anon Coward), I'll give you the TL;DR preview, I'll simply list the caveats, without much explaing. Please remember we are talking about SERVERS. For companies. Not desktops, not laptops, not gaming rigs, not servers for a home lab. Some people seem to forget that...
0.) If your workload needs anything Intel propiertary, like transactional memory, or some other weird stuff, then game over. By the same token, if you need something AMD exclusive, like the niffty feature that encrypts a VM's RAM so that not even the VMM can see it, then go AMD.
1.) Does your Workload (usually legacy) depend on single threaded performance? In that case, is likely that intel will be faster, but do a side by side trial and benchmark, as that will depend on the performance loss for the Spectre patches for both intel and AMD.
3.) Is your software licensed by core, not socket? (If it is FOSS, then the answer is, no, move on) Then, you need the configuration which achives the desired performance with the MINIMUM numer of cores, even if the Processor that achieves that is supper-expensive... The Software tends to be ORDERS OF MAGNITUDE MORE EXPENSIVE then the SERVERS (I am looking at you Oracle). So, again benchmark, but remember AMD has more cores, per socket, but intel tends to have more single thread performance, so, again, side by side trial and benchmark.
4.) Does your workload has problems with NUMA (also, mostly legacy)? In that case, certainly, intel. (with the use of "core-complexes" XEN has more "NUMA-ness" than intel, by a long shot. Even a single socket XEN processor is NUMA).
5.) If you are doing virtualization or cloud, does your VMM/Cloud SW support AMD properly? If not, is it worth it for your organization to add a new VMM to your stable on top of a new server chip? Like it or not, not all VMMs/Cloud SW support AMD yet, I hope this changes in the near future.
6.) If you need to LIVE migrate VMs from the old kit to the new kit and back, then the kit has to be either all intel, or all AMD. (migrating turned off machines is easier, the problem is LIVE migrating)
7.) Are the server makers in your approved provider list selling AMD servers? Is it worth it for you as a tech to go through the process with finance/purchases to add a new server provider?
8.) Are the server makers that sell AMD servers providing good service in your country? Are they providing good service to your company? Service varies by region and account.
9.) Are the server providers that deal with you providing AMD servers in the Form-Factor your project demands. I've seen mostly blades thus far. If you need 4U with many 2.5"disks, or 1U, or some other format, is it made by the server sellers your company approves?
Finally, a server is more than the microprocessor it uses, and actually, the microprocessor is a small part of the cost. Maybe the server seller is trying to get this project for strategic reasons, or increase their footprint in your company, or hit the sales goal so that they collect the bonus and pay their yatch. In the end, your project needs servers that cover it's needs in terms of security, performance, form factor and support, amog other things. If after all these hurdles are cleared, for whatever reason, a server seller is offering you a less expensive intel server that covers your needs, go for it. If the less expensive one is AMD, go for it
As you see, none of this is in AMD's control, it will get better over time, and none of this is due to evil machinations from intel.
I hope this is usseful to you, and hope I've got a new "friend" on /.
Let me know what you think.
PS: For Grammar nazis, I bet my english is better than your spanish, AAAAND I also can communicate in french*, not too shabby for an electronics engineer.
*used to be able to speak, read and write french, but after 15 years of no practice, I forgot most of it.
Supongo que los asesoro para el branding la misma gente a la que se le ocurrio el mitsubishi "pajero"
For the english speaker, Poco in spanish means little, while pajero means wanker.
my post in english says:
It seems to be very little, better not buy it.
I guess that they got branding advice from the same people that came up with the mitsubishi pajero.
If Intel completely reset the processor to its previous state after a failed speculative execution, then the processor would be immune to all speculative execution attacks--even speculative execution attacks that haven't been discovered yet. But it doesn't appear that that's what Intel is doing. Instead, their strategy appears to be to design a separate patch for each different speculative execution attack after that attack is discovered.
What did you expect? It takes about four years to engineer an new microprocessor architecture. More if you are starting from scratch. I am 99% certain that right now, there is a team of Intel Microarchitects designing such microprocessor from scratch, we may as well see the fruits of their labour in something like six years (maybe more, if there are more slowdownd in the Fab side of things). But in the meantime, Intel needs something to sell, otherwise, the company would go bankrupt before that new microprocessor design is out the door, so, we end up with this.
And before you think about going to some other manufacturer (AMD, Arm, IBM Power, Spark, VIA), remember that their processors are also suceptible to Spectre type attacks, so do a cost/benefit business case analysis before taking the plunge...
PS: I favour AMD for servers (since this article is about servers) but there are many caveats beyond AMD's control that make it an uphill battle to use their chips. This has nothing to do with obscure machinations from Intel like in the early '00s, and fortunately will get better in the future, but, again, the thing is, do a detailed cost/benefit analysis before taking the plunge.
In this decreases lifespan, but mantains, or actually increases, the number of healthy years in that decreased lifespan, is actually a welcome development.
No, is not made of tachions. The way standards work, is that drafts are circulated, and barring any major weirdass cornercases, what is ratified is pretty close, if not exactly equal tot he draft.
Remember all those "Pre-n" wifi APs at the end of last decade? Similar thing.
Therefore support for TLS 1.3 (as described in the draft) was added to firefox 52, and the draft was ratified into official standard in Aug 2018. No tachions involved.
If you want slide out landscape keyboard, you should take a look at the MotoZ play with a Livermorium Slide-out keyboard MotoMod.
I do not know if it supports LineageOS, but the ShaterShield screen will add enhanced durability to the package...
If you mean unlocked BOOTLOADER, as in, I can install custom ROMs, than Sony is the one to beat. Just be careful of the specific model you select.
If by uunlocked you mean "I can stick a SIMcard from any opearator and the thing will work", then a Nokia 6 or higher is the best Android for that job.
Onorable mention for the MotoZ with Shatershield + MotoMods.
As I stated on my original post, I use Firefox ESR 60 on a mac. And Firefox on my android (KeyOne).
At home I use 9.9.9.9, 8.8.8.8 and 208.67.222.222 since I have better things to do than to set up my Synology to be my DNS server.
But when I travel, I use public wifi whenever I can get it, be that my hotel, the training centers were I teach, or, god forbid, a hipster coffee shop. And many of those need a captive portal to autenticate to the Wifi, and that depends on using the Network's DNS servers. So, I configured an "Automatic" setting on the network locales of my mac to handle those cases.
So, as a user of Firefox, I am not happy with this. I am capable enough to configure my DNS settings (or, if push comes to shove, set up a DNS from scratch, not even touching my nas).
So thank you for the inconvenience mozilla. I hope you guys enjoy the backslash when hipsters start to realize that they can not connect to the net in their favourite hipster watering spot because they can not get to the captive portals...
At least, the guys who use Mozilla in corporate networks will get this assinine setting turned off in group policies... as for the rest of us, a quick google and a trip to about::settings shal suffice