Ultimately, the "2019" iMac models are an internal upgrade that provide faster performance and graphics performance than the "Mid-2017" iMac models that came before them. Otherwise, the two lines are quite similar.
That said... users have figured out how to shoe-horn NVMe drives in the Mac Pro, offering higher-tier performance and much better prices. Unfortunately, no one has taken the time to compile a list, so the known so far are: Samsung 960, Samsung 970 Pro, Toshiba XG3, and Crucial P1. Samsung released a firmware fix for certain models as well, including the 970 Pro.
In December 2013, Apple released a new cylindrical Mac Pro (colloquially called the "trash can Mac Pro"). Apple said it offered twice the overall performance of the first generation while taking up less than one-eighth the volume. It had up to a 12-core Xeon E5 processor, dual AMD FirePro D series GPUs, PCIe-based flash storage, and an HDMI port. Thunderbolt 2 ports brought updated wired connectivity and support for six Thunderbolt Displays. Reviews initially were generally positive, with caveats. Limitations of the cylindrical design prevented Apple from upgrading the cylindrical Mac Pro with more powerful hardware.
Original marketing materials for the Mac Pro generally referred to the middle-of-the-line model with 2 × dual-core 2.66 GHz processors. Previously, Apple featured the base model with the words "starting at" or "from" when describing the pricing, but the online US Apple Store listed the "Mac Pro at $2499", the price for the mid-range model. The system could be configured at US$2299, much more comparable with the former base-model dual-core G5 at US$1999, although offering considerably more processing power. Post revision, the default configurations for the Mac Pro includes one quad-core Xeon 3500 at 2.66 GHz or two quad-core Xeon 5500s at 2.26 GHz each. Like its predecessor, the Power Mac G5, the pre-2013 Mac Pro was Apple's only desktop with standard expansion slots for graphics adapters and other expansion cards.
In April 2018, Apple confirmed that a redesigned Mac Pro would be released in 2019 to replace the 2013 model. Apple announced this new Mac Pro on June 3, 2019 at the World Wide Developers Conference. It returns to a tower design similar to the Power Mac G5 in 2003 and the first-generation model in 2006. The design also includes a new thermal architecture with three impeller fans, which promises to prevent the computer from having to throttle the processor so that it can always run at its peak performance level. The RAM is expandable to 1.5 TB using twelve 128 GB DIMMs. It can be configured with up to two AMD Radeon Pro GPUs, based on RDNA 1 architecture, which come in a custom MPX module, which are fanless and use the chassis's cooling system. Apple's Afterburner card is a custom add-on, which adds hardware acceleration for ProRes codecs. Similar to the second generation, the cover can be removed to access the internals, which features eight PCIe slots for expansion, making this the first Mac with six or more expansion slots since the Power Macintosh 9600 in 1997. It can also be purchased with wheels and in a rack mount configuration. Feet and wheels are not stated by Apple to be user-replaceable and require sending the machine to an Apple Store or authorized service provider, though tear-downs show the feet are simply screwed on. It was announced alongside the Pro Display XDR, a 6K display with the same finish and lattice pattern.
Since November 2020, Apple debuted its first Macs equipped with M1 chips, notably in the 13-inch MacBook Pro, MacBook Air, Mac mini, iMac, and iPad Pro models. This silicon chip benefits in performance from the previous Intel-based Mac, but it also means that the steps to erasing each drive differ slightly.
The Apple MacBook Pro 14 with a base M1 Pro scores exceptionally well in Geekbench 5. It has remarkable single- and multi-thread performance, similar to the top-end M1 Max in the Apple MacBook Pro 16 (2021) but with a slightly lower multi-thread score. The GPU compute score is decent, in the same ballpark as the NVIDIA GeForce GTX 1650 in the HP Pavilion Gaming Laptop 15 (2021).
But what if Intel only offers base models of its Xeon Scalable CPUs and then allows customers to buy the extra features they need and enable them by using a software update? This is what SDSi enables Intel to do. Other use cases include literal upgrades of certain features as they become needed and/or repurposing existing machines. For example, if a data center needs to reconfigure CPUs in terms of clocks and TDPs, it would be able to buy that capability without changing servers or CPUs.
At long last, there's an official and finalized specification for the next generation of High Bandwidth Memory. JEDEC Solid State Technology Association, the industry group that develops open standards for microelectronics, announced the publication of the HBM3 specification, which nearly doubles the bandwidth of HBM2E. It also increase the maximum package capacity.So what are we looking at here? The HBM3 specification calls for a doubling (compared to HBM2) of the per-pin data rate to 6.4 gigabits per second (Gb/s), which works out to 819 gigabytes per second (GB/s) per device.To put those figures into perspective, HBM2 has a per-pin transfer rate of 3.2Gb/s equating to 410GB/s of bandwidth, while HBM2E pushes a little further with a 3.65Gb/s data rate and 460GB/s of bandwidth. So HBM3 effectively doubles the bandwidth of HBM2, and offers around 78 percent more bandwidth than HBM2E.What paved the way for the massive increase is a doubling of the independent memory channels from eight (HBM2) to 16 (HBM3). And with two pseudo channels per channel, HBM3 virtually supports 32 channels.Once again, the use of die stacking pushes capacities further. HBM3 supports 4-high, 8-high, and 12-high TSV stacks, and could expand to a 16-high TSV stack design in the future. Accordingly, it supports a wide range of densities from 8Gb to 32Gb per memory layer. That translates to device densities ranging from 4GB (4-high, 8Gb) all the way to 64GB (16-high, 32Gb). Initially, however, JEDEC says first-gen devices will be based on a 16Gb memory layer design."With its enhanced performance and reliability attributes, HBM3 will enable new applications requiring tremendous memory bandwidth and capacity," said Barry Wagner, Director of Technical Marketing at NVIDIA and JEDEC HBM Subcommittee Chair.There's little-to-no chance you'll see HBM3 in NVIDIA's Ada Lovelace or AMD's RDNA 3 solutions for consumers. AMD dabbled with HBM on some of its prior graphics cards for gaming, but GDDR solutions are cheaper to implement. Instead, HBM3 will find its way to the data center.SK Hynix pretty much said as much last year when it flexed 24GB of HBM3 at 819GB/s, which can transmit 163 Full HD 1080p movies at 5GB each in just one second. SK Hynix at the time indicated the primary destination will be high-performance computer (HPC) clients and machine learning (ML) platforms.
The CXL 1.1 specification supports three protocols: the mandatory CXL.io (for storage devices), CXL.cache for cache coherency (for accelerators), and CXL.memory for memory coherency (for memory expansion devices). From a performance point of view, a CXL-compliant device will have access to 64 GB/s of bandwidth in each direction (128 GB/s in total) when plugged into a a PCIe 5.0 x16 slot.PCIe 5.0 speeds are more than enough for upcoming 3D NAND-based SSDs to leave Intel's current Optane DC drives behind in terms of sequential read/write speeds, so unless Intel releases PCIe 5.0 Optane DC SSDs, its existing Optane DC SSDs will lose their appeal when next-gen server platforms emerge. We know that more PCIe Gen 4-based Optane DC drives are incoming, but we haven't seen any signs of PCIe 5.0 Optane DC SSDs.Meanwhile, CXL.memory-supporting memory expansion devices with their low latency provide serious competition to proprietary Intel Optane Persistent Memory modules in terms of performance. Of course, PMem modules plugged into memory slots could still offer higher bandwidth than PCIe/CXL-based memory accelerators (due to the higher number of channels and higher data transfer rates). But these non-volatile DIMMs are still not as fast as standard memory modules, so they will find themselves between a rock (faster DRAMs) and a hard place (cheaper memory on PCIe/CXL expansion devices). 2b1af7f3a8