Today is the second part of the review of Intel’s Arc B580 Limited Edition, which was unintentional but necessary in this form. Today you’ll find out everything about the board and the components used, as well as other details that you’re unlikely to find in every gaming review or YouTube video. We will shed light on where Intel may have copied from AMD or NVIDIA, question the frequency spectrum in the audio measurement when it comes to the fans and their control and, of course, thermally analyze the entire board, including a necessary conversion. After careful consideration, however, I have prefaced all of this with an important foreword, because I will also have to write about why I didn’t pack everything into one article straight away, but why there had to be a sequel today.
How not to deal with hardware and your colleagues
I have already spoken with Intel and contacted the colleague from a German media outlet in question (unfortunately, I haven’t received a response yet). Even though the card was in rotation—since not everyone could get a card directly from the manufacturer in the first round—you can still expect some respect for the product and consideration for the colleagues who test it afterward. The improper assembly of the card cost me two whole days of work, as I had to redo a significant portion of my benchmarks. This also resulted in two lost days for paid lab tests. Workstation tasks always take longer, after all.
I don’t want to judge the article hidden behind a paywall (thank you to a reader for granting me access), but when I read it, I found no meaningful reference to the PCB or the cooler. It barely touched on anything; they might as well have left the card as it was. As a principle, I only disassemble cards after I’ve tested everything. That’s why it took me so long to come to the conclusion that I can’t trust anyone without a proven reputation anymore. Aside from the mechanical damage to the card (tool marks) and improper handling, the use of completely unsuitable thermal paste and torn thermal pads resulted in thermal disadvantages that I can’t just blame on Intel, as it clearly wasn’t their fault. And since I was provoked, I collected the paste using tools and even tested it according to standards—because there was plenty of it to work with. This paste ranks just slightly above the hand cream I tested years ago, for anyone who remembers. More on that later, as it will form the “story within the story” for today.
I will therefore repeatedly refer to these circumstances throughout the article, substantiating them with measurement data and the repair that has already been carried out. This is also to exonerate Intel because you can’t just leave it at that. Sure, the issue with the rotation isn’t ideal either, but I think there’s a solution for that in the future. However, I had promised Intel to publish the benchmarks before Christmas. The charts in the first article already include the results with the repaired card, and I essentially had to write the article twice and ultimately split it due to time constraints. Thanks to a certain colleague—though, given that you knew exactly who would receive this card directly from and back to you—my initial reaction wasn’t exactly fit for print. Precisely for these reasons, I deliberately ignored the delivery note, which indicated I should confirm receipt of the sent card.
Anyone who feels capable of disassembling a card should have the original or at least equivalent materials—whether it’s PTM, good thermal paste, or soft pads—on hand. Not to mention good Torx screwdrivers. I hereby publicly offer to gift the colleague a year’s supply of quality thermal paste, a large strip of PTM and soft pads for the VRM and memory, as well as a set of gloves. I will cover the shipping and material costs personally, entirely without publicity, ulterior motives, or a delivery note.
But necessity can also be turned into a virtue, and with a bit of distance, one can unemotionally comment on how to do things better. Since I already had the card open once, and the soft pads (semi-putties) Intel used weren’t quite as fresh after multiple uses, today I’ll also provide a paste and pad mod with the original PCM8500 (OEM), Maxtor CTG10, and Thermal Grizzly TG Putty Pro to address the hotspots as an additional contribution to the content.
The Intel Arc B580 in detail
I’ll save myself the majority of redundant content, as much has already been written in great detail by my colleagues. So just a refresher on the essentials: The Intel Arc B580 Limited Edition is a powerful graphics card based on the BMG-G21 chip of the Battlemage architecture. It has 20 Xe cores with a total of 2,560 shader units and 20 ray tracing units. The graphics processor works with a base clock of 1,700 MHz and reaches up to 2,670 MHz in boost mode. In practice, a clock rate of 2,850 MHz is often observed under load. The card is equipped with 12 GB GDDR6 memory, which is connected via a 192-bit interface. The effective memory clock is 19.2 Gbps, which enables a memory bandwidth of 456 GB/s.
An 8-pin PCIe connector is required for the power supply and the maximum power consumption is 190 watts. The Arc B580 Limited Edition offers a variety of connectivity options, including one HDMI 2.1 port and three DisplayPort 2.1 ports. One of the DisplayPorts supports Ultra-High Bit Rate 13.5 (54 Gbps), while the other two offer UHBR 10 (40 Gbps). With these specifications, the Intel Arc B580 Limited Edition is an attractive option for gamers and content creators who value high performance and modern connection standards. The dual-slot card is around 28 cm long, 11.5 cm high and weighs just under 790 g.
The obligatory GPU-Z screen will conclude this introduction, because this time it’s straight down to the nitty-gritty and thus also the teardown from page two. If you’re going to do it, do it properly.
- 1 - Introduction and important foreword
- 2 - Teardown: cooler design, material analysis and fan in detail
- 3 - Teardown: circuit board, topology and components
- 4 - Thermal paste from the hell
- 5 - Pad or putty? That will be interesting!
- 6 - Paste vs. pad and putty, high-res thermography, noise
- 7 - Power consumption and load peaks
- 8 - Summary and conclusion
40 Antworten
Kommentar
Lade neue Kommentare
Veteran
1
Veteran
Urgestein
Urgestein
1
Veteran
Urgestein
Veteran
1
Veteran
Neuling
Veteran
1
Urgestein
Urgestein
1
Mitglied
Veteran
Alle Kommentare lesen unter igor´sLAB Community →