Nvidia have had a firm grip on the GPU market for some time. Having not had any real competition. They’ve promoted and command huge mind share.
However AMD have been upping their game over the last almost two years. Enough to push Nvidia to fast track its GPU roadmap. Phasing out some of its initial RTX offerings and putting out tweaked “Super” versions of their cards as AMD were bringing out their initial NAVI GPUs.
Nvidia then filling out its low to mid range GPU offerings as AMD announced theirs. What was interesting was that Nvidia felt no shame as what they brought to market was upgraded GTX10 series cards.
Running counter to their own marketing that RTX was the future.
So Nvidia have been playing cat and mouse with AMD for many months. Countering AMD’s offerings just before they’re actually released. Basing their products on AMD’s announcements.
However with this latest round, Level1Tech’s posed an interesting question. “Did AMD purposefully blindside Nvidia through last minute software revisions and pricing to improve the 5600XT.”.
As they’ve said, the only issue is how many customers would actually apply a BIOS update to get that 10%+ performance boost.
Arguably all of this has been great for the consumer. Especially since the bad old days of cryptocurrency mining where GPUs were stupidly expensive. If there was any stock to be had.
On pure price / performance, AMD have a solid range. However Nvidia have done significant work differentiating their GPUs based on use. This does come at a premium though.
Nvidia keep and maintain strong relationships with software companies. Working with not just games but also utilities and productivity.
Everything from ray tracing Minecraft. CUDA technology incorporated into Adobe Creative suite and machine learning acceleration.
My only question would be, “Does the new crop of GPUs really justify the prices?”. Nvidia are arguably artificially keeping prices high as a holdover from the crypto mining days.
All said and done. The PC consumer is in a very good place at the start of this new decade. There hasn’t been this level of competition and choice in some time.
So Mac Pro day has come and gone. Its taken awhile to get my head around just what exactly the 2019 Mac Pro does and where it fits.
Workstation. HEDT. Theres a lot of categories and even more user expectations.
On the face of it. The Mac Pro, especially the base configuration, is an extraordinary bad value. The single core and multi core benchmarks are pretty average. Factor in the price of the base config in your country and, yeah. Its looks very hard to justify.
Apple have done themselves a disservice with the Pro name. Across the product range its applied to mean various things.
With iPads, its a nebulous range of additional features that don’t really distinguish the iPad Pro all that much from the more mundane models. There’s just the horrendous price.
Similarly, the Pro in the iPhone 11 range denoting the premium line. Again, the price to features being a little nebulous. Just expensive.
The Macbooks are a lot clearer. The Macbook Pros are the most capable. Especially the 2019 16 inch Macbook Pro. However they are priced as a mainstream device. The balance of processor, memory and storage for the relative price is bearable.
This diverges more with the iMacs. The iMac Pro is kitted out similarly to the Mac Pro with workstation grade internals. Running a Xeon processor and ECC memory. It offers a good value proposition in terms of processor, memory and storage, even for its steep base price.
Its biggest issue is that it requires an Apple technician to do upgrades.
As an aside, the elephant in the room for the iMac Pro is “Does it have a future?”. Theres been no update for some time.
Enter the Mac Pro. Its clearly a workstation and the price paid is for all the engineering thats gone into its design.
As Quinn’s video shows and iFixIt’s rating confirms. The Mac Pro is totally user serviceable. Initial testing is confirming that its capable of handling loads easily without sounding like a jet engine.
Viewed from this angle. As a purely commercial workstation. The Mac Pro makes sense.
The problem is that configurability is table stakes in the PC world. For general computing and hobbyists, its hard to justify. Its something thats a given with high end desktops.
Part of the overall problem is on Intel. The glacial pace of the Xeon roadmap and supporting chip set features. Its hard not to look at what AMD Ryzen and Threadripper bring to the table.
Performance is very much on par with the chip specs. Single core scores are tied to clock speed and multi-core scales out with the number of cores on tap.
There are still other trade offs. For example, Nvidia have done significant work to accelerate machine learning. However Apple have intentionally disallowing support for their GPUs.
The Mac Pro is competitive in key areas like video editing. Some features like the Afterburner card and custom dual GPUs are a clear advantage.
However its harder to say beyond that. There are as many options at any and all price points if macOS isn’t mandatory. All of which are flexible.
While Apple forces its customers to make sacrifices at all price points.
The Mac Pro comes at a huge premium for its flexibility. Hard to justify beyond the small market that can. Which is a shame as it still leaves a huge unfulfilled customer market.
Courtesy of Techpowerup. AMD are pretty much hitting every CPU bracket just in time for Christmas.
High end desktop – Ryzen 9 3950X. An AM4 socket CPU with 16 core / 32 threads. A clock-speed of 3.50 GHz base, with 4.70 GHz maximum boost frequency, at 105 W TDP. Priced at $749US.
Note the retail box package won’t come with a cooling solution. AMD recommends at least a 240 mm x 140 mm AIO liquid CPU cooler.
Available from November 25, 2019.
Budget desktop – Athlon 3000G is an unlocked 2 core / 4 thread CPU. Base clock is 3.5Ghz. Its running a Vega 3 integrated GPU. Not overly powerful but for $49US its a solid option for a NAS, HTPC or low end desktop.
Available from November 19, 2019.
High end work station / server – 3rd generation Threadripper. These are for the really good boys and girls. Requiring a new sTRX4 socket but compatible with existing TR4 coolers. That cooler will need to be a beefy one as AMD are rating these new CPUs at 280 watts TDP.
These will need the new TRX40 chipset motherboards. Supports quad channel DDR4 memory up to 2Tb. ECC if its really needed. It also brings PCIe4 support as well. Note TRX40 is not compatible with older Threadrippers.
3960X is a 16 core / 32 thread CPU for $1399US. Running at 3.80 GHz base with up to 4.50 GHz maximum boost.
3970X is a 32 core / 64 thread CPU for $1999US. Running at 3.70 GHz base, and 4.50 GHz maximum boost frequency.
Available November 25, 2019.
Neato bonus feature. Zen 2 ECO mode. This will be a configurable TDP control in the Ryzen Master software. Any processor higher than a Ryzen 5 3500 can be capped at 65 watts on the fly. A no brainer feature really. A little like cylinder deactivation in a car.
A good overview of some of the new Surface hardware that was announced.
A little more on whats under the hood of the Surface Pro X, Surface Pro 7 and Laptop 3.
As handy as the Pro 7 looks. The Laptop 3 15″ looks more useful to me. The bigger screen and discrete GPU are big pluses. Along with a design allowing service. That goes a long way towards peace of mind in the future.
What was announced was all a little ho hum. Even the presenters seemed a little non plussed although it did get better. I did only watch the edited presentation though.
World’s worst kept secret. Google Pixel 4.
Pricing here in Australia is ok. Although for just a little more theres the iPhone 11.
Its an alright update. Apple seems to have jumped ahead in terms of lens but it’ll be interesting to see if the Pixel keeps its photo crown.
The jury is out on gesture controls. Its nice to not see a notch. Is it enough to make a mark even in the Android market? The 3a. I’m just saying.
Being where I am, we only got the revised Nest Mini and Nest Wifi. Good iterative updates. The Nest Wifi inheriting smart speaker functionality is a no brainer.
Unfortunately Australia doesn’t get the Pixelbook Go. This is positioned more as a regular device where last years ambitious tablet went down in flames.
I don’t know that Google weren’t wrong in what they wanted the tablet to be. Its just that it was half baked, going by the reviews. Microsoft are having a tilt at this with the Surface Pro X. The Pro X may go over better as Microsoft are pushing aggressively and it is running Windows.
Thats all folks.
Addendum. As of time of writing. Google and retailers are doing $150 redeemable vouchers on the Pixel 4.
The Surface Neo has floated around Microsoft’s design studio for some time. This looks like Microsoft’s wildest bet. The demonstration was extremely hand wavy but showed the use cases.
Actual hardware details are scant. The promo video showing that it runs Intel so compatibility should be no problem. A customised OS to handle the 2 screens, Windows 10X. Apparently its pronounced, Windows ten ten?
The interesting part is the Windows UI customisation to the magnetic keyboard. The LCD panel switching to a track pad mode or a minimised screen / Touch Bar depending on where the keyboard sits.
Given the small size and internals, the battery life will be key. This looks pitched at a similar market to the Surface Go. The Go has been popular but one of the key knocks has been poor battery life.
Its an intriguing device nonetheless. Able to meet a wide variety of use cases while still being super portable.
The issue will be, will it find an audience. Where the Go neatly filled the low end, affordable portable niche. The Neo is likely going to fall between the Go and the regular Pro.
Again, price will be a major factor. Especially when the customer hears that the keyboard and Surface pen costs extra. (Pure conjecture but thats how the Surface Pro has rolled.)
Then for the almost “One more thing” moment. Don’t call it a comeback. Panos Panay jokingly had something left to announce.
It wasn’t presented as a phone.
Given how mature the smart phone market is. Microsoft having learnt a lesson from their first go round has partnered with Google. The special sauce being the work to allow Android to work with a dual screen UI.
It’ll be interesting to see just how much work was done. I’ve no doubt that Microsoft will have a uniform user experience within its own apps. But how soon does the UI fall apart as disparate Android apps are loaded.
The over arching issue here is development. Theres disparate versions of Windows on different architectures and then theres Android. I know Microsoft has been working hard on platform agnostic development tools.
Successfully so. At this point, they’ve managed to claw back a lot of developer mindshare. Delivering great open technology and tools. Even building Windows out as a developer OS. Can I get a high five for WSL!
However they’re all very different paradigms. Some like Windows 10X and the dual screen Android UI still in development.
To be successful, these products need to find an audience. Microsoft themselves have productivity covered. But a healthy App Store is needed. Even if its just to convert a sale.
Microsoft has always been a leader in development tools. So much so, that Windows always led in the desktop OS wars. Winning developer mindshare allowing it to dominate business and consumer markets with software alone.
However the computing landscape is very different today. The ubiquitous computing platform isn’t a desktop computer. <Billie Eilish>Duh</Billie Eilish>
At the core, each target platform has its own particulars and APIs / tool sets. Microsoft has clearly said they want developer input. The problem will be just how much work is involved in delivering to all these different devices and will it be worth it.
Users want a good selection of apps. Developers want a large paying user base to sell to. The issue is kick starting an App Store which has neither.
Third party developers will be faced with deciding on what to deliver to. Its a similar problem that Apple has been trying to address. Although from a very different angle.
Even Apple has struggled. When they have the advantage of a more uniform product set and significantly more control over those products. Along with one of the biggest App stores and largest paying user bases focused almost exclusively on the iPhone.
With Swift and unified APIs, Apple are hoping to make delivering macOS and iPadOS simpler and easier. Enough so as to entice iOS developers to broaden these smaller markets.
Microsoft has yet to address any of this in a significant way. Maybe the next Build ? If they can develop this into a platform then they can leverage those sweet sweet subscription services.