* Posts by BinkyTheMagicPaperclip

72 posts • joined 11 May 2012

Page:

Hypervisor indecisive? Today's contenders from yesterday's Hipsters

BinkyTheMagicPaperclip

This article is fundamentally wrong

Hyper V is a type 1 (bare metal) hypervisor, with cut down Windows derived components (not full fat Windows) on top of it.

Xen, and also Xen server, is a type 1 hypervisor. It uses a paravirtualised OS (designed to be Xen aware) as a domain 0. Dom0 manages access to devices (technically you can run devices off a stub domain, but that's adding complexity), it may also run Qemu which provides emulated devices only - not CPU emulation. The Dom0 can be Linux, NetBSD, or Solaris. For Xenserver, dom0 is a version of Linux.

Why would you use NetBSD as a dom0? Well, Xen is GPL2, but NetBSD is BSD licensed and there's been a fair bit of work performed to create custom embedded NetBSD kernels for specific purposes. Just be aware if you're using PCI passthrough that there's a fair few Linuxisms you'll have to work around, and that passthrough of graphics cards is currently non functional.

There are then DomUs (guest domains) that range from fully hardware virtualised (potentially Xen unaware), to fully paravirtualised (completely Xen aware). The drivers used within these domains can then again either be hardware virtualised, paravirtualised, or in-between (can improve performance).

KVM is a type 2 hypervisor. It runs alongside Linux (and very badly under FreeBSD, with hacks and the Linux compatibility kernel module. Don't bother) as a qemu accelerator. Qemu performs all the instruction translation with optional KVM assistance, it also provides the emulated devices. KVM also comes bundled with vfio on Linux, the most functional PCI passthrough support, particularly in the case of graphics passthrough. These are all separate components. Vfio will work with qemu and without KVM.

FreeBSD has bhyve, its own type 2 hypervisor which works with the FreeBSD kernel. This is quite new, as of FreeBSD 10.0.

Jails are something entirely different.

8
0

HORDES OF CLING-ONS menace UK.gov IT estate as special WinXP support ends

BinkyTheMagicPaperclip

Re: Really?!

Nope, read the article :

'NHS Scotland, like NHS England, is not responsible for leading or forcing IT strategy at a grass-roots level.'

The trusts know. NHS England doesn't.

2
0

JavaScript CPU cache snooper tells crooks EVERYTHING you do online

BinkyTheMagicPaperclip

The second time I'm covered for still running an old platform!

Main system :

Adjacent row reading in memory exploit : not vulnerable, ECC memory, Core2Quad based.

This cache exploit : Core 2 Quad, not vulnerable.

Firewall : pentium 3.

Course, there's a tablet (baytrail) and a shortly to be commissioned Ivy Bridge server. So I'm not in the clear, and at work there's both a load of old, not vulnerable stuff and new i5/i7 vulnerable kit..

0
0

Rip up your AMD obits: Gaming, VR, embedded chips to lift biz out of the red by 2016, allegedly

BinkyTheMagicPaperclip

Re: Faulure to understand

20%? I'm willing to bet it'll be more than that. AMD can promise what they like but have basically been nowhere on the desktop for years other than the cheap integrated CPU/GPU systems at the low end for which they're definitely worthwhile, and the fact FX chipsets support ECC.

I see no mention of virtualisation support - an area where AMD could innovate, but it's Intel that have been driving that market. No mention of TSX. No mention of an alternative to XenGT - AMD could easily release a high end desktop or server/workstation CPU with powerful inbuilt GPU for sharing amongst VMs both for higher end graphics and compute. It's an obvious hole in Intel's lineup (Xeon's don't include an integrated GPU in most circumstances). No mention of HSA on FX processors - so it's basically a low end gimmick AMD aren't serious about.

They're not getting used in half decent laptops or tablets - Baytrail is eating the latter market.

Then there's the GPUs, where theoretically they're better in open source but in practice NVidia is a preferred option if you're not using NetBSD/OpenBSD/FreeBSD without a binary driver. On Windows the NVidia drivers still appear to be better and they have stereoscopic 3D built in rather than needing a third party product. NVidia supports GSync which at the moment is clearly better than AMD's alternative. Against that, there's the fact that NVidia can't be relied upon to keep any sort of openness (witness the disabling of CUDA when an AMD card is present in the same system, and the recent refusal of their non professional cards to run under hypervisors)

2
1

Relax, it's just Ubuntu 15.04. AARGH! IT'S FULL OF SYSTEMD!!!

BinkyTheMagicPaperclip

Re: systemd? Do not want.

Salix is much better than Slackware - all the BSDs have integrated package management, and its omission in Slackware by default is just too painful. The use of lilo instead of grub is unusual these days, as is the fact it boots without an initrd, but the use of mbootpack enables even complex Xen configurations to boot.

I did have to terminate Networkmanager with extreme prejudice - I'm sure it must be possible to get a bridge working reliably on startup using it, but life is too short. It's also on a system which just runs a lot of VMs and will never run X, or add/remove physical network adapters (virtual is a different matter).

0
2

Windows 10 Device Guard: Microsoft's effort to keep malware off PCs

BinkyTheMagicPaperclip

x86 has four protection rings, of which commonly only 2 are used (with the honourable excelption of the horrific and weird IOPL DLLs in OS/2 that run at ring 2). x64 has less rings and operates a little differently

0
0
BinkyTheMagicPaperclip

This is opening a wriggly can of worms

The support for this will probably be limited to a small selection of hardware. The main issue here is not that it's a bad idea to use an IOMMU (it's probably a good idea), the issue is that everything will be running under a hypervisor.

It will doubtless be Windows 8's Hyper V with improvements

The issues with this are

1) Speed. It will be (slightly) slower. A worthwhile tradeoff, perhaps.

2) Drivers. A VM is not the same as real hardware. It may break some drivers or degrade their functionality (particularly graphics drivers)

3) Communication between mini VM and the wider world. If it needs to do this, presumably it's via a network card and will require two IP addresses. If via the main Windows VM, that's an attack surface. It'd have to be an SR IOV compliant network adapter (more expensive), as otherwise multiple cables are require, surely.

4) Cross expansion card communication. An IOMMU only protects communication between a VM and a card/memory that is not assigned to it if the PCI-e root port the card is attached to supports ACS. Otherwise one VM with one card assigned, can write to the memory space of another card in another VM, when they both share the same (non ACS protected) root port. ACS is not supported on plenty of implementations

5) BIOS. You'll need a new system simply because the quality of consumer BIOSes for VTd/IOMMU is pathetic and manufacturers will not fix it because Windows historically hasn't needed it aka everything from Asus and 'we don't support Linux'

I'll be interested to see what happens when it's run under an existing Hypervisor - my Windows 8 installation already runs under a hypervisor (Xen), using an IOMMU (passthrough of network, graphics cards to the VM), on a Core2 CPU no less (not that I would recommend this)

The security record for hypervisors isn't bad, but there has been information leakage/denials of service between VMs and to the hypervisor itself. It's not a magic bullet.

The bright side of this is that running a decent hypervisor on commodity hardware may become substantially cheaper!

1
0
BinkyTheMagicPaperclip

Re: Why do I get a bad feeling about this...

..although, this will naturally be a version of Windows 8's Hyper-V, with additions. Therefore it will require a 64 bit CPU with second level address translation, which is indeed Nehalem onwards for Intel.

See http://blogs.msdn.com/b/uk_faculty_connection/archive/2012/10/24/hyper-v-list-of-slat-capable-cpus-for-hosts.aspx

1
0
BinkyTheMagicPaperclip

Re: Why do I get a bad feeling about this...

VT-d is a chipset technology, not a processor technology. From Nehalem onwards the memory controller was integrated into the CPU package meaning that the lines became increasingly blurred.

VT-d works fine (for varying values of 'fine') on Core 2 chipsets provided it is the right chipset (X38, X48, S3200/S3210, most of the Q chipsets) and the BIOS has it enabled with a bug free implementation (in reality this means nothing from Asus will work, most Intel boards will, plus Supermicro, some DFI IIRC)

1
1

If hypervisor is commodity, why is VMware still on top?

BinkyTheMagicPaperclip

Sort of.. Xen/Xenserver are different

Xenserver is built on Xen, and is pretty much a turnkey product as long as you like the way it works, not unlike ESXi in many ways. It's also now open source (except for, yes, VDI bits and a few other things)

Xen itself is a tad different - it's the base hypervisor, and sometimes a bit of an arse to get going depending on what you're doing with it. You can tweak it at a fairly low level, though, which has its advantages, as is the fact it can work with just about any (x86, arm) Linux distrubution, NetBSD x86 (does a few things better than Linux, others not so well) and Solaris (never tried).

However, it's a solid product and the design is mostly good. KVM is better and faster at a few things (PCI passthrough, for one), but doesn't seem to hang together quite as well - it's basically an (optional) qemu accelerator with a (rather good) vfio addon for hardware passthrough, rather than a bare metal hypervisor with a coupled host OS that drives devices.

Just don't be tempted to use xen -unstable if your favourite feature of the day isn't around yet, you're likely to regret it. Specific, tested xen releases are fine, and if included in your distribution much easier than compiling from source.

ESXi is bloody fantastic *if* all your hardware is supported, your network is perfect, nothing needs to be tweaked and you don't need to also use the glass console of the ESXi server. The problem comes when it doesn't work, the management tools decide not to play ball, or the next version with a killer feature is released, and it decides to kill off support for your motherboard/network card/raid controller for no particular reason other than because they can't be arsed to support it. In that case, just buy the new hardware - do *not* try and hack it to work.

For work usage, I'd consider ESXi on nice certified hardware, Xenserver or at a pinch RHEL with KVM. For home, if VMWare Player/Workstation isn't sufficient and ESXi is overkill, I'd go for KVM first, then Xen.

13
0

Nvidia's GTX 900 cards lock out open-source Linux devs yet again

BinkyTheMagicPaperclip

(then again..) Re: Nvidia's Unix support may be a tad limited

..on the other hand, I may have to give Nvidia their due.

My configuration hadn't worked, but finally stick in the latest Nvidia driver, released only days ago, and it Only Bloody Works! Wasn't included on the fix list, but it's now ok. Fantastic.

0
0
BinkyTheMagicPaperclip

Re: yeah, sounds familiar

Glad it works for you, but can you clarify which driver release you're using? You may get a large shock if you haven't updated for 9+ months and try to now - the driver used to work in passthrough. Now it does not, because Nvidia decided it should be that way. It definitely broke Xen and KVM; I thought it did the same for VMWare's products.

There are patches widely available to KVM/Qemu to hide that they're running under a hypervisor. The same patches are not as prevalent for Xen and non-existent for VMWare for obvious reasons.

From NVidia's point of view, passthrough is only supported under Xenserver (not Xen), RHEL 6 with KVM and ESXi, for certain Quadro and GRID cards only (not all). Consumer cards are not supported.

2
0
BinkyTheMagicPaperclip

Nvidia's Unix support may be a tad limited

You may be lucky under Windows, but hit a Unix issue and you may not be so fortunate- unless you're a big customer.

Say you want to get graphics passthrough to a VM working - something a bit leading edge you can't do with a consumer card without tricks, because Nvidia have deliberately disabled it in their drivers (reason no 1 for open source drivers, although they're currently not good enough). So, you buy a professional card (modding may also work), try again and it still fails. The error is clearly in the drivers because it works with one card and one operating system, but not in another operating system.

You report it to Nvidia and it gets rejected until tested on a validated platform. Faff around building a valid platform, that your hardware is ok and one card is fine and the other does not work. Response? Please provide a business case as we are short of resources.

I shouldn't be too mean because Nvidia support have been reasonably helpful, but this is a situation where the hardware is capable, the open source drivers are not, Nvidia support won't release info to update the open source drivers or support professional hardware with official drivers on an enterprise operating system with server hardware components.

(This is actually for FreeBSD using binary drivers. The official binary drivers work with passthrough under Linux and Windows. This only serves to amplify the main point however - or even more so under other BSDs where there are *only* open source drivers)

Course, the main problem is that Nvidia is still the best option out there. Intel isn't too bad (if slow) but their driver quality is variable, and more importantly they don't sell discrete graphics cards any more. Consumer CPUs? built in Intel graphics. Xeon? Nada. AMDs hardware is generally not as capable and the driver quality is sub par. Matrox, if you must, for a couple of the cards only works if it's the primary GPU on your system.

It's a huge, complex mess with entrenched interests. I don't see it changing soon. I did get my system working, using a more expensive, larger and power hungry card, but its hassle I could have done without.

3
0

Tinder Plus charges oldies MORE to ogle young hotties' pics

BinkyTheMagicPaperclip

This is not new

OKCupid, also owned by match.com (as tinder are too), runs age based pricing for premium membership. Fortunately you can reduce your age, buy the membership more cheaply, then raise your age again..

Tinder is a waste of time, anyway. Even if you're within standard cookie cutter dating parameters (young, single, monogamous, straight and want to have children eventually) its insistence on not having a home location for each person and instead matching on where someone was five days ago makes it utterly useless. Gave up on it very quickly as a total waste of time.

3
0

Elite:Dangerous goes TITSUP

BinkyTheMagicPaperclip

Re: I've been thinking about playing....

I can't speak for E:D because I never play online games, however the experience should be vastly better with an analogue joystick. I'd recommend the Thrustmaster T16000M (thirty to forty quid) as it has accurate hall effect sensors.

Playing oolite (free, Elite like game) with an analogue joystick to dock is so much easier than the keyboard it's a pleasure - you can match rotation speed with the station precisely and guide your ship gracefully in.

4
0
BinkyTheMagicPaperclip

mmm.. Schadenfreude

I thought I'd moved on, downloaded some other space sims and RPGs and decided I had better things to do than worry about offline mode being dropped from E:D

well, hello, schadenfreude! Looks like I'm not willing to forgive Braben for his shitty last minute behaviour, yet. I should feel more sorry for the players than I do, but currently I'm laughing at Frontier's past statements that an offline game lets you cheat, but an online one does not..

14
6

El Reg's festive dating app guide, Pt. 2: The FEMALE perspective

BinkyTheMagicPaperclip

online dating does work

..particularly if your lifestyle or requirements are in any way beyond a standard opposite sex relationship with a day job and 2.4 kids. I meet a lot of new people, and online dating is actually pretty damn good (comparatively) in terms of finding someone available, of a vaguely suitable age, attractive enough, where you fancy each other. You can filter out and get a yes/no answer comparatively quickly.

Meeting people is not difficult. Meeting the right available people is.

0
0

Lonely this Xmas? Nerds, n00bs and no-hopers' guide to dating apps, Pt.1

BinkyTheMagicPaperclip

OKCupid is good, POF is alright

People will actually look at your profile (as a man mostly looking for women, this is not common practice on all dating sites) and I've had decent chats, a number of dates and made some friends.

POF is a tad horrific at times, but I have had dates and casual encounters off it (sometimes they wanted more, but I didn't with them. Other times it's vice versa) so I can't really complain.

I didn't get on well with match.com when I tried last, and guardian soulmates was a waste of time.

0
0
BinkyTheMagicPaperclip

Re: Rugged?????

It's got multiple meanings. I guess in some cases it could mean straight acting but in my experience it either means 'I want to meet without my partner knowing' or 'I'm queer and my partner does not know' (*)

Whilst I'm not inundated, it is not unknown as a bisexual man with a dating profile stating I'm almost entirely looking for women, to receive an e-mail from guys who are absolutely, 100% straight, yet still want to suck me off.. I guess I'm just irresistible *cough*.

They almost always start with 'hi' and it's never worth replying as you know what's coming next. Openly queer guys are much nicer.

(*) technically I suppose it could be the other way around, but it's about as rare as two lesbians looking for a man - it does happen, but it's terribly uncommon.

1
0

Time to bag a tablet bargain this Chrimbo – Brit stocks are high

BinkyTheMagicPaperclip

use a Windows tablet?

Windows Bay Trail tablets, yes, because it's real Windows 8.1 on an Intel processor. Typically the screen resolution is 800p landscape, though

0
0

Sporty in all but name: Peugeot 308 e-THP 110

BinkyTheMagicPaperclip

Wake me up when it's less boring or does more mpg

Given the choice of a Focus or this I would definitely go for the Focus.

45mpg is not impressive, and matches up with the other small engined cars I've driven. Either buy something with decent performance and crap mpg, or excellent mpg and lacklustre performance. A middle of the road car is no use to anyone.

A half decent diesel will do 45mpg when run ragged. Driven carefully it'll do 60+...

4
1

Linux systemd dev says open source is 'SICK', kernel community 'awful'

BinkyTheMagicPaperclip

Re: I understand the poor guy but probably public whining is not a good idea

Creating a compatibility layer just to get software to work is a pain. There is a rapidly developing culture that ignores diversity of platforms and is segmenting operating systems into clearly defined niches.

By all means create a pluggable startup system, but don't tie a desktop environment to it and specifically to Linux.

It used to be that operating systems had some similarity. Now there's not even any consensus about how to configure network interfaces (ifconfig on BSD, ipconfig/netsh on Windows and fuck knows on Linux - depends on the distribution).

Now there's systemd that's Linux only and increasingly foists dependencies on a load of software, Xen is practically Linux only even if in theory it works with a NetBSD and Solaris dom0, FreeBSD are creating their own hypervisor, OpenBSD their own mail server and web server.

There are often reasonably good reasons for this, but it's still a colossal pain in the arse that skills are increasingly non transferable.

Not to mention that Linux is increasingly deprecating swathes of hardware and deciding historical interfaces are troublesome because they inconveniently get in the way of new shinies..

7
0

Xen security bug, you say? Amazon readies GLORIOUS GLOBAL CLOUD REBOOT

BinkyTheMagicPaperclip

Re: First BASHing

They can - live migration is supported on Xen provided there's shared storage available so potentially there should be almost zero downtime.

However, the update from EC2 says this isn't possible, and until the details are released on 1st October we can't know if this is reasonable or not.

It does say that if your database is replicated correctly it's possible to reboot ahead of time and avoid downtime..

2
0

AMD claims record with latest overclock-happy FX Series chips, again

BinkyTheMagicPaperclip

Re: It isn't eight cores and it's still embarassingly slow

Go on then, point me at some decent benchmarks.

The dirty tricks you refer to relate to some sub optimal tuning in Intel's compilers. Intel were fined by the FTC and had to fix the issues.

Any non Intel compiler benchmarks I can find show the FX processors' performance to be inconsistent at best. Add in the fact that AMD is second choice for optimising on virtualisation software and the reality is that AMD has to produce a significantly better chip than Intel in all areas, at the same price point to make it worth considering. They managed that years ago, but failed to capitalise on their advantage whilst Intel was still struggling with hot P4s.

0
0
BinkyTheMagicPaperclip

It isn't eight cores and it's still embarassingly slow

It's little better than hyperthreading and sometimes not even that. It has four actual cores with two integer execution units in them. Comparing with an Intel Core i5 at the same price the i5 tends to have 30% better single threaded performance. The FX series have 10-30% multithreaded performance advantage in a limited selection of benchmarks.

The AMD A series are worth buying if you want a fast enough processor with inbuilt graphics fast enough for modern games. Everything else? Don't bother. At the higher end (FX9590 vs i7 4790) the comparison is just embarassing. AMDs fastest chip is slower in almost every benchmark whilst consuming 2x the power, having slower memory speed and usually no USB3.

AMD need to knock 50 quid off their CPU prices at the low end and 70 at the high end to be even vaguely equal. Even then it's not a convincing win. Roll on the next generation, this one is a lemon.

0
6

Ofcom will not probe lesbian lizard snog in new Dr Who series

BinkyTheMagicPaperclip

Re: Priorities

Practically none. If you kiss your pet the same way you kiss a human you fancy then I worry about a) your pet or b) your unsatisfied partner

Personally I thought they're one of the best things to happen in Who in ages. At least they're comprehensively getting rid of screaming girlies, so common in classic Who

22
3

LibreSSL crypto library leaps from OpenBSD to Linux, OS X, more

BinkyTheMagicPaperclip

Re: Pity they dropped Windows support

Windows support is coming - they're currently working on it

0
0

AMD tops processor evolution with new mobile Kaveri chippery

BinkyTheMagicPaperclip

Yawn, again

I'm sure it might be ok at the low end, and honestly like AMD's concept of a combined CPU/APU to speed up tasks - if implemented correctly it could indeed by fast.

Problem is, where is the code? There's at least one bit of custom code that speeds up part of one application but that's about it. To deliver on AMD's promises there needs to be pervasive, tangible speedups baked into operating systems rather than being vaguely comparable with Intel.

There's nothing available for Windows or Linux as far as I can see. AMD does well at cryptography and gaming but lags at almost everything else.

That's not necessarily a bad thing, if this is a reasonably performing low end chipset. Unfortunately AMD is trying to position this as a new architecture. Until they deliver on HSA code it's just marketing fluff.

Typical AMD : it's a good idea on paper, but execution is lacking.

0
2
BinkyTheMagicPaperclip
Thumb Down

Re: It seems to me...

Show me the benchmarks demonstrating the FX processors destroying Intel's chips.

Believe me, I'd quite like to run an FX processor, especially as it supports ECC RAM. I've looked at a lot of benchmarks and AMD almost always loses. It doesn't really have 8 decent cores and is embarrassingly slower at compilation compared to the i7 whilst running hotter.

That's not to mention that the support chipsets are substandard too. Very few motherboards support PCI-e 3.0 now and as far as I can make out, all the exciting virtualization progress is being made by Intel (don't argue 'PCI-e 3.0 isn't needed', systems should last for years and future proofing is important as is bandwidth on <= 4x physical cards)

AMD hasn't been competitive for years, except at the low end, where their built in GPU is much better than the Intel offering.

1
2

OpenBSD founder wants to bin buggy OpenSSL library, launches fork

BinkyTheMagicPaperclip

Re: Madness is doing the same thing and expecting a different outcome

The advantage of C is portability. C++ compiler quality is *still* variable on different platforms and there's the name decoration issue. Yes, it's possible to work around it, but C is a tad simpler.

Additionally pretty much all existing operating systems are written in C (although there are exceptions such as L4, and certain components written in C++. Windows is written in C, but some parts are realistically only usable from C++ or .NET).

For most libraries and non low level operating system code I would use C++, but in the specific case of libraries such as OpenSSL that may be running on embedded systems with fairly severe resource restrictions this may not be realistic.

1
0
BinkyTheMagicPaperclip

Not a good idea. I don't know Bronek's coding skills, but hacking cryptography should be left to the very small number of people that are good at both coding and cryptography.

0
0

Running OpenSSL? Patch now to fix CRITICAL bug

BinkyTheMagicPaperclip

Change freeze hit in February. Note there's already an ICMP errata from March. Media are covenient but in reality you'll need to update via. STABLE build . You should be checking the errata page often

0
0

Wii got it WRONG: How do you solve a problem like Nintendo?

BinkyTheMagicPaperclip

Don't know what you're talking about, Nintendo have a great current gen console

It's visually impressive, fast enough, innovative, uses social features well and has a great back catalogue.

What, Wii U? Hell no, I'm talking about the 3DS XL! Fantastic handheld and the 3D is used well on many occasions. Also loving indie games in the marketplace.

3
0
BinkyTheMagicPaperclip

You can with a link cable, yes. It is most excellent in Pacman Vs. Also there's Zelda : Four Swords where you can team up with three other GBA players!

1
0

Apple Mac Pro: It's a death star, not a nappy bin, OK?

BinkyTheMagicPaperclip

ingenious, but also rather lacking a point

I can't help but agree with others and think the machine would look an awful lot better with a transparent cover and a couple of lights (despite the fact I don't generally go for lighting and my dad snarkily asked 'why do computers need lights? Are they scared of the dark?')

I can understand the wisdom of an external breakout box but think it should be an AND, not an OR. I have a Power Mac G4; it's still substantially smaller than my main workstation capable of holding EATX motherboards and a load of storage. This is the top level product and shiny design should take a backseat to sheer all out power and storage.

Perhaps few people will want to have more than 12 cores, but it's nice to have the option and more cores necessarily means lower clock speeds to fit in Intel's 150W thermal envelope (they appear to be pushing things a bit with Ivy Bridge EX)

Previous Mac Pros were extremely expandable, fast, had the option to fit as yet unreleased processors (won't happen in this case) and were still maintainable and quite pretty.

Hell, if they insist on going to the smaller form factor why not go for 'borg cube'. Looks good on the desk, also arrange it so it fits in say 5U with the ability to mount more than one horizontally. It'd solve the absence of the XServe too.

0
0

You’re NOT fired: The story of Amstrad’s amazing CPC 464

BinkyTheMagicPaperclip

You're being sniffy over semantics. The database capability in Mallard BASIC was excellent and it was a particularly decent language to code in (I used the PCW version which was somewhat inferior at screen handling and any graphics required assembly(*), but allowed for much larger programs by loading code in and out of memory via the RAM drive)

Also, I rather suspect you're replying to Roland Perry.

. He knows a lot about this.

(*) I'm not counting GSX

1
0

For Windows guest - KVM or XEN and which distro for host?

BinkyTheMagicPaperclip

It does work, incidentally!

Just beware you'll have issues. On Windows Catalyst must be installed manually by selecting the driver, then installing the CCC MSI. Running the install always results in a blue screen.

It is fast enough to run games. You'll have to fiddle to get the best disk performance - use the virtio drivers. If anyone is trying to run older OS, be aware that KVM/Qemu creates a VM which is quite similar to a Q35 chipset, but with differences. With ancient OS you may need to use the Qemu 'pc' architecture (440LX). It may also be necessary to use a CPU type of qemu64 or qemu32 in some cases rather than 'host' or enabling KVM

Remember that the VFIO or pcistub driver is separate from KVM. Passthrough works without it. KVM only provides acceleration, which is usually (but not always) faster.

The virtual PC that KVM, Qemu and Xen create is similar to a real PC but it is not the same. OS/2, for instance, does Weird Shit(TM) on install (to be precise, non mainstream OS may tickle VMs in a way that's entirely valid on real hardware but freaks out the VM).

If I was doing this professionally, I'd use Xen and stick to a released version ideally without VGA passthrough. Xenserver is now free and a nice piece of kit.

What I really should have done is to buy a dual Xeon system, with a quadro and run Xen. What can I say, I'm waiting for Haswell-E before upgrading and spending lots of money. In the meantime I'm running an unusual Core2Quad system with a 6950 (pre Nehalem VTd works, but has no interrupt remapping).

1
0
BinkyTheMagicPaperclip

KVM is your best option, if you insist on your requirements

Whilst I'd suggest Linux on Windows is your best option overall, if you insist on your requirements KVM is the best option.

Xen is a great piece of software, but in the region of VGA passthrough it is decidedly inferior to KVM. Yl not get any support if it does not work.

Xen supports only Quadro devices for reliable passthrough. KVM supports AMD and Intel, but you will need a very recent Linux kernel (3.12+) and patches, plus a recent, patched Qemu. Google 'vfio VGA reset'

Be very careful with your hardware choices and read around the subject first. Also, read the motherboard manual cover to cover before purchase. My motherboard, for instance, supports graphics cards at 4x PCIe speed in only one slot - all others are limited to 1x by the chipset.

Also be careful of USB passthrough. In Linux/Windows it mostly works. In BSD it does not - you have to do a VTd passthrough rather than a single device passthrough in Qemu. You cannot usually pass through a single port due to iommu groups. USB works through allocating a pool of resources to a certain number of ports. Again, on my motherboard I have five USB ports. These can be passed through in two iommu groups - so 2/3 or similar? Wrong! 1/4, or 4/1!

I'm doing this because I'm enjoying fiddling with low level Linux, virtualisation and I'm stubborn. It's still leading edge stuff and you can expect to encounter pain. Be very familiar with iommu groups, PCIe bridges, FLR, VT-d and if your cards support various types of reset before you buy any hardware.

1
0

HARD ONES: Three new PC games that are BLOODY DIFFICULT

BinkyTheMagicPaperclip

Re: Nethack

Nethack is only difficult when you don't have the experience, although it takes a lot of dying and/or reading the handbook to gain the experience. Slash 'EM or ADOM are more tricky. If you fancy a giggle, though, I recommend DoomRL - it's fun and reasonably complex without needing to memorise a lot..

I'd suggest Dwarf Fortress. Can be terribly tricky and there's a whole O'Reilly book about it.

0
0
BinkyTheMagicPaperclip

It's not the 80s, and things have moved on.

Think back to the 80s - Jet Set Willy 1 and 2. Bright, addictive, varied and punishingly difficult to complete. Once you died it was yet another 10-20 seconds to return to the first room and then a number of minutes to reach your last progress point and die. Again.

It didn't make it any more interesting that few people were able/bothered enough to complete it.

Now examine the modern equivalent, VVVVVV. Bright, addictive, varied and punishingly difficult to complete. Once you died you were placed straight back into the action. Your progress was not lost and this made it all the more interesting.

Once complete, it was interesting to complete again and unlocked further challenges.

Some of do actually play modern FPSes for the story, scenary and a bit of excitement which is why I'm a fan of Jedi Knight, Half Life etc and couldn't care much less about Quake 3 and all subsequent deathmatch variants.The fact a lot of saves are used and levels finished on single health points is not the issue. If the game was properly designed 'save scumming' would make little difference - poor tactics and repeated saves can get you through marginal battles. Make the battles *require* skill or tactics and the use of a savegame stops being an issue.

1
0

AMD's 'Revolution' will be televised ... if its CPU-GPU frankenchip Kaveri is a hit

BinkyTheMagicPaperclip

Re: Dull and slow

For basic productivity : probably not. For browsing, quite possibly. For recent games, definitely (moot point, don't play much bang up to date).

For virtualisation and compiling, yes. I've looked at the benchmarks. AMD's performance vs Intel is embarassing in single threaded performance and it doesn't fare well in compilation benchmarks. The difference is most pronounced under Windows but there's still a noteable gap under Linux.

I'd quite like to buy an AMD system, but the facts don't support what I want to be the case. There's some blather on Gentoo forums about the improvement with the latest GCC etc ,but until I see verification I'm putting this down as someone wanting to brag about their system.

0
0
BinkyTheMagicPaperclip

Dull and slow

It's still way slower than a 4770K unless you're using built in graphics and playing games. There's no AMD provided software to show the potential of this supposedly new architecture. Frankly it looks like another acknowledgement of the fact they can't match Intel in CPU power, even when they overclock their chips and peg them at 220W TDP.

I would like to give AMD a try, really, but unless the pricing is keen and the intention is a lower end gaming desktop I can't see the point. If you move to something like an FX8350, currently the only advantage is that it can support ECC somewhat cheaper than an Intel platform.

With an offering like this, Intel has no incentive to increase core counts or decrease prices for their chips, especially the consumer level ones.

0
3

Anatomy of a 22-year-old X Window bug: Get root with newly uncovered flaw

BinkyTheMagicPaperclip

Re: Well on of my distributions had the fix in yesterday

That's not the same bug. The latest bug was patched 2.5 days ago - check the diff history for bdfread.c

0
0

You're fired: Lord Sugar offloads faded PC builder Viglen to XMA

BinkyTheMagicPaperclip

Re: So that makes three computer businesses flushed...

Nonsense - he actually did rather well. The quality of his av kit may have been a bit variable but let the owners copy tapes cheaply, which is what they really wanted it for.

PC wise the 1512/1640 were a revelation at the time - working and undercutting everyone else. The PPC series had a horrid screen but again worked and was very cheap for the time. Amstrad's in house developed CPC, PCW series were lovely pieces of kit, as was the somewhat niche NC100.

I don't think many would deny he improved the Sinclair line somewhat, and had enough sense to dump the QL.

After the 80s Amstrad did somewhat go into decline, but during the 80s it really succeeded.

0
0

No sign of Half-Life 3 but how about FOURTEEN Steam Machine makers?

BinkyTheMagicPaperclip

Re: @adniam

SteamOS works on 'regular Linux' provided that's only very recent Ubuntu. With anything else (Debian, for instance)l it's an arse to get running as it uses very recent versions of libc and other libraries.

SteamOS is not a regular OS. It's designed to be linked up to the TV and not tinkered with. Given the amount of fettling I do on my main system that does have its attractions.

Having said that, the Linux catalogue is still too limited; even Valve's software is largely incomplete. One of the reasons I sorted a system that runs KVM/Xen with hardware passthrough to a graphics adapter was partly for fun and development and also because I could run Windows' Steam..

0
0

'Weird' OBJECT, PROPELLED by its OWN JETS, spotted beyond Mars orbit by Hubble

BinkyTheMagicPaperclip

Noooo! It's a Vorticon invasion!

Fools - Yorp is not a group of people, it's a small green single eyed Martian alien! Mostly harmless, but there are aggressive examples. Beware of other species such as the technically advanced and warmongering Vorticons!

Ring the US and get Billy Blaze on the line, we need Commander Keen to save us all!

(I think the Yorps disappeared after the first couple of releases of Commander Keen - perhaps they're stranded on that asteroid)

1
0

Microsoft warns of post-April zero day hack bonanza on Windows XP

BinkyTheMagicPaperclip

OEM has always been about the motherboard

what XP will activate on or the Microsoft phone registration staff will give you a code for is irrelevant. OEM Windows is tied to the motherboard. If the motherboard dies it must be replaced by a comparable motherboard (same one or generation, not an upgrade).

the only differences between OEM and retail are the motherboard lock, lack of telephone support and the inability to resell the license (although selling the motherboard/PC with the license is ok). It's all in the license terms, it's just that most people choose to ignore them..

0
0

PlayStation 4 is FreeBSD inside

BinkyTheMagicPaperclip

Re: *looks at Eadon and laughs*

yep, hotplug does work on *BSD although in some cases it's not out of the box.

BSD tends to be a lot more predictable and better documented than Linux, and the install programs are pretty solid (I've had a lot of grief with Linux distros trying to be too clever then failing badly when installing on something vaguely unusual)

driver support in Linux is better especially for brand new graphics cards. application support depends on the BSD variety - NetBSD is very portable but IIRC you're restricted to older versions of Skype and Firefox rather than Chromium. OpenBSD focuses on openness and security, so you won't be seeing ZFS any time soon.

For a Linux user FreeBSD is probably more familiar and friendlier than the other BSDs - you might find NetBSD shockingly bare bones. On the other hand if you look closer OpenBSD is remarkably coherent and well thought through.

for all occasional hassle I've had with *BSD there are equivalent issues with Debian, Arch, Ubuntu etc especially when you want to tell them to stop being 'clever' and do exactly what you want (insisting on the use of the network notification stuff, for instance, rather than 'here is my nic, here is my bridge, do not mess with them - that's the user's job')

4
0

Review: Renault Zoe electric car

BinkyTheMagicPaperclip

Re: Range issue

More than you think, and for the hard of thinking : the recharge time is too long. 200 miles in one go is much rarer than a few 70 mile trips.

0
1
BinkyTheMagicPaperclip

Re: 75 miles? hahahahahahahaha

We're clearly not going to agree, but I will concede that most journeys are under 75 miles - most of the ones I'm thinking of are.

The problem is the recharge time. If I take a typical Saturday's activities :

Route 1 - 13 miles each way. Travel time half an hour (car), 1:11 (bus).+25 minutes extra wasted waiting time

Route 2 - 23 miles each way. Travel time half an hour (car), 1:30 (average - train, walk, bus). Have to leave half an hour earlier than in car.

Total=72 miles. Total time wasted if public transport used : 3 hours, 45 minutes + leaving half an hour early.

Heaven forfend I actually fancy nipping to the shops as well! Alternatively, how about going to my parents (65 miles). You can get there, but not back...

No, I don't have time to waste half an hour recharging at a charging station, and the places I'm travelling to definitely don't have recharge facilities. I am not willing to concede that my expectations are in any way unrealistic - they can be easily met by the most basic non electric car available.

Five fat blokes in a VW Up with a sofa is not a common occurrence for most people. Four fat blokes without the sofa might be - if the VW Up couldn't fit four moderately sized people (perhaps with a degree of discomfort in the back) I'd criticise that too..

There are plenty of small city cars that are a little underpowered for the motorway - but they can manage it if necessary.

A 200 mile range would be reasonable - all the local journeys would be covered, plus occasional day trips into the countryside. Holidays would be viable with overnight charging, too.

0
0

Page:

Forums