京东黑号检测方法分享

qq316107934: 可以试下这个链接： https://item.m.jd.com/product/100038651615.html

如果能到付款界面 则没问题，提示账号火爆则说明某种程度上你的号被京东标记了。

找了 7 个朋友做测试，3 个轻度购物的用户可以顺利到付款界面，4 个会逛张大妈重度电子数码用户但是没薅羊毛行为的无法购买提示活动火爆。

我是 plus 会员，从来不薅羊毛也不拆单之类的，就正常买，也黑号了。

找客服人机，无法告知原因，只会重复明天再试。

狗东就这么玩吧，明年不续了。

时效性：2025-03-10

Read More

怎么取消预览,Windows 11 ，当鼠标指向任务栏程序的时候，会弹出预览，很烦人。

mfsw:

试了以下方法，无效：1 组策略 用户配置 → 管理模板 → 开始菜单和任务栏 找到 关闭任务栏缩略图 选项，双击后选择 已启用 并应用设置

2 注册表 定位到路径：HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Advanced ExtendedUIHoverTime 击打开后设置数值数据为 9000

定位到路径：HKEY_CURRENT_USER\SOFTWARE\Microsoft\Windows\CurrentVersion\Explorer\Taskband NumThumbnails 双击后设置数值数据为 0

诸位大佬搭救一下，感谢。

Read More

零跑 B10 预售 1 小时订单破 15010 辆，朱江明称是品牌史上最多的一次

IT之家 3 月 10 日消息，零跑汽车官方宣布，其全新车型零跑 B10 在开启预售仅 1 小时后，订单量便突破了 15010 辆，刷新了品牌历史记录。零跑创始人在朋友圈表示，这是零跑产品发布以来，一小时内订单量最多的一次。

零跑 B10 于今日晚间正式开启预售，共推出 5 款车型，售价区间为 10.98-13.98 万元。这一价格区间在市场中极具竞争力，尤其是其全系标配智驾功能。其中，售价 12.98 万元的 510 智驾版车型配备了激光雷达高阶智驾功能，打破了国内激光雷达智驾车型的价格底线。

在智能化配置方面，零跑 B10 配备悬浮式仪表盘和 14.6 英寸的 2K 中控屏，入门车型即搭载 8155 骁龙芯片，高配车型则配备 8295 旗舰芯片，并采用 Leapmotor OS 4.0 Plus 交互系统。车机系统接入通义千问和 DeepSeek 大模型，支持 AI 语音控制，能够精准理解指令并执行相应的车控功能。

在智能驾驶方面，零跑 B10 全系均为智驾版车型，标配自动泊车功能。特别是激光雷达智驾版，支持端到端大模型智驾上车，可实现高速及城区自动领航辅助。

IT之家注意到，作为一款家用 SUV，零跑 B10  主副驾座椅靠背均可直接 180° 放平，全车配备 22 处储物空间，后排座椅放平后，车内可扩展至 1415L 容积，满足家庭大储物空间的需求。动力及续航方面，零跑 B10 全系采用单电机后驱，零百加速最快可达 6.8 秒，提供 510km 和 600km 两种续航里程选择。

相关阅读：

Read More

AMD EPYC 9845 Makes For A Persuasive Upgrade With Performance & Energy Efficiency

With the new AMD EPYC 9005 processors there are SKUs up to 500 Watt with the likes of the EPYC 9965 flagship at 192 cores for Turin Dense cores or 128 Turin classic cores with the EPYC 9755. But for those looking at upgrading from an existing EPYC 9004 series server and bound by the motherboard BIOS support and/or cooling/power capacity, 400 Watts is a sweet spot. Many of the existing platforms designed for EPYC 9004 Bergamo/Genoa(X) and now extended for EPYC 9005 Turin are limited to a 400 Watt TDP. With the prior AMD EPYC 9655 testing I have already shown off the great Zen 5 uplift when maintaining the same core counts as Zen 4, but even sticking to 400 Watts at the top-end is room for more. The EPYC 9845 is AMD's top-end SKU for 400 Watts or less that allows for 160 dense cores (320 threads) per socket compared to the 128 core EPYC 9754 Bergamo. Effectively the same power level and 25% more -- and better (Zen 5C) -- cores. Plus with EPYC Turin supporting the new AMD P-State CPU frequency scaling driver there is greater headroom in optimizing for power efficiency if so desired. Here is a look at how the AMD EPYC 9845 delivers a great leap to performance and power efficiency for those looking at a surprisingly robust upgrade from prior generation EPYC 9004.

AMD EPYC 9845 CPU

For those limited by the server platform's electrical characteristics and/or cooling capacity or even at the limits for your rack's available power, the AMD EPYC 9845 can provide a very compelling option while remaining within the 400 Watt TDP envelope. The AMD EPYC 9845 provides 160 cores / 320 threads of the Turin Dense variety, 2.1GHz base clock, 3.25GHz all-core boost speed, and 3.7GHz maximum boost clock. The EPYC 9845 has a default TDP of 390 Watts and a custom TDP range from 320 to 400 Watts. There is 320MB of L3 cache for this processor, 12 channel DDR5-6000 support, and other features in common with the rest of the AMD EPYC 9005 line-up.

AMD EPYC 9845 bottom

The EPYC 9845 can provide a very meaningful upgrade compared to the Bergamo top-end at 128 Zen 4C cores / 256 threads. Besides 25% more cores and all the great uplift from the Zen 5 architecture, the EPYC 9845 has a 115MHz higher all-core boost speed and 600MHz higher maximum boost clock while also having the 320MB cache compared to 256MB with the Bergamo processor. The EPYC 9754 has a 360 Watt default TDP but can be increased up to a 400 Watt cTDP.

AMD Titanite CPU heatsink for 400 Watt CPUs

The EPYC 9845 can run in a 2P server configuration as well for providing up to 320 cores / 640 threads per server while again striking below the 400 Watt threshold. For the purposes of today's testing I was comparing the EPYC 9754 Bergamo and EPYC 9845 Turin processors in a single socket (1P) configuration. These tests were done on the AMD Titanite reference server platform that AMD used for Genoa/Bergamo. With a simple BIOS upgrade, I was able to switch to using Turin CPUs... With this upgrade story, I stuck to using the DDR5-4800 DIMMs as well for both the Bergamo and Turin Dense benchmarking. If wanting an even greater upgrade path, going from DDR5-4800 to DDR5-6000 yields very nice gains with Turin as shown in other articles.

AMD EPYC 9845 vs. EPYC 9754

The AMD Titanite server was running with air cooling, 12 x 64GB DDR5-4800 Samsung memory, and was running Ubuntu 24.10 for fresh benchmarks of an up-to-date Linux stack. I was also using the Linux 6.13 kernel for the latest AMD P-State driver support. The CPU configurations tested for this article included:

EPYC 9754 - The 128-core Bergamo processor with ACPI CPUFreq and the performance governor. Only AMD EPYC 9005 "Turin" and future server processors are defaulting to using AMD P-State with newer versions of the Linux kernel.

EPYC 9845 - The 160-core Turin Dense processor at its default with the amd-pstate-epp driver on Linux 6.13 and the performance CPU frequency scaling governor.

EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem - If wanting the 160 core upgrade path but also desiring the best energy efficiency, this run represents that with switching over to the powersave governor on AMD P-State while also switching over to the "power" Energy Performance Preference (EPP) and the Balanced Memory Power Profile. See the prior article AMD EPYC Turin Power Profile Selection Impact On Performance & Efficiency for more information on the tuning options and impact on power/performance.

EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem - Similar to the above run but keeping to the "performance" EPP value.

EPYC 9845 - acpi-cpufreq performance - For those on an older version of the Linux kernel without amd_pstate driver support or on a server without ACPI CPPC as needed by AMD P-State or just wanting a 1:1 comparison to Bergamo, this run is using the EPYC 9845 while switching back to the ACPI CPUFreq performance configuration as used by the EPYC 9754.

AMD EPYC 9845 Benchmarks vs. EPYC 9754 Linux Comparison

The CPU power consumption, CPU thermals, and performance-per-Watt were all looked at as well as part of this 128-core EPYC 9754 vs. 160-core EPYC 9845 comparison for the 400 Watt server space.

OpenVINO benchmark with settings of Model: Face Detection FP16-INT8, Device: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.

Kicking off this benchmarking bout with OpenVINO for some AI coverage, the performance upgrade from Bergamo to Turin Dense while not going above the 400 Watt threshold was dramatic right away. The switch from EPYC 9754 to EPYC 9845 yielded 1.5x the performance with OpenVINO! That's for just 25% more cores and not even upgrading from the existing DDR5-4800 server memory. Just a BIOS upgrade and CPU swap to the existing Genoa/Bergamo 2U server platform.

OpenVINO benchmark with settings of Model: Face Detection FP16-INT8, Device: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.
OpenVINO benchmark with settings of Model: Face Detection FP16-INT8, Device: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.
OpenVINO benchmark with settings of Model: Face Detection FP16-INT8, Device: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.

Even more of a kicker was that the EPYC 9845 was consuming significantly less power than the EPYC 9754. The AMD P-State and Power Profile tuning also paid off with dropping the CPU power consumption even more.

OpenVINO benchmark with settings of Model: Age Gender Recognition Retail 0013 FP16-INT8, Device: CPU. EPYC 9845 was the fastest.
OpenVINO benchmark with settings of Model: Age Gender Recognition Retail 0013 FP16-INT8, Device: CPU. EPYC 9754 was the fastest.

Helping out the EPYC 9845 deliver such superb OpenVINO AI performance is the full 512-bit data path found with the EPYC Zen 5 server processors. Switching to DDR5-6000/DDR5-6400 memory would also help in squeezing even more performance out of this existing server platform.

OpenVINO benchmark with settings of Model: Face Detection Retail FP16-INT8, Device: CPU. EPYC 9845 was the fastest.
OpenVINO benchmark with settings of Model: Face Detection Retail FP16-INT8, Device: CPU. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.
OpenVINO benchmark with settings of Model: Face Detection Retail FP16-INT8, Device: CPU. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.
OpenVINO benchmark with settings of Model: Face Detection Retail FP16-INT8, Device: CPU. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.

The EPYC 9845 was consuming 81% the CPU power on average of the EPYC 9754 while engaging the amd-pstate powersave and EPP tuning led to 75% the power use of that Bergamo flagship while still achieving similar EPYC 9845 performance levels.

Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Llama-3.1-Tulu-3-8B-Q8_0, Test: Prompt Processing 512. EPYC 9845 - acpi-cpufreq performance was the fastest.
Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Llama-3.1-Tulu-3-8B-Q8_0, Test: Prompt Processing 512. EPYC 9845 - acpi-cpufreq performance was the fastest.
Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Llama-3.1-Tulu-3-8B-Q8_0, Test: Prompt Processing 512. EPYC 9845 - acpi-cpufreq performance was the fastest.

With Llama.cpp the EPYC 9845 was delivering 1.84x the performance of the EPYC 9754 prior generation processor while consuming less CPU power.

Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Llama-3.1-Tulu-3-8B-Q8_0, Test: Prompt Processing 512. EPYC 9845 - acpi-cpufreq performance was the fastest.

The reduced CPU power consumption also correlated to lower temperatures too on this AMD Titanite server.

Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: granite-3.0-3b-a800m-instruct-Q8_0, Test: Prompt Processing 2048. EPYC 9845 was the fastest.
Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: granite-3.0-3b-a800m-instruct-Q8_0, Test: Prompt Processing 2048. EPYC 9845 was the fastest.
Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: granite-3.0-3b-a800m-instruct-Q8_0, Test: Prompt Processing 2048. EPYC 9845 was the fastest.
Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Mistral-7B-Instruct-v0.3-Q8_0, Test: Prompt Processing 2048. EPYC 9845 - acpi-cpufreq performance was the fastest.
Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Mistral-7B-Instruct-v0.3-Q8_0, Test: Prompt Processing 2048. EPYC 9845 - acpi-cpufreq performance was the fastest.
Llama.cpp benchmark with settings of Backend: CPU BLAS, Model: Mistral-7B-Instruct-v0.3-Q8_0, Test: Prompt Processing 2048. EPYC 9845 - acpi-cpufreq performance was the fastest.

The EPYC 9845 provides for a very heavy hitting upgrade path for existing EPYC 9004 servers (or other 400W TDP limited products) that top out at a 400 Watt TDP to achieve not only much greater performance but very meaningful efficiency improvements.

Blender benchmark with settings of Blend File: Classroom, Compute: CPU-Only. EPYC 9845 was the fastest.
Blender benchmark with settings of Blend File: Classroom, Compute: CPU-Only. EPYC 9845 was the fastest.
Blender benchmark with settings of Blend File: Classroom, Compute: CPU-Only. EPYC 9845 was the fastest.
Blender benchmark with settings of Blend File: Barbershop, Compute: CPU-Only. EPYC 9845 - acpi-cpufreq performance was the fastest.
Blender benchmark with settings of Blend File: Barbershop, Compute: CPU-Only. EPYC 9845 - acpi-cpufreq performance was the fastest.
Blender benchmark with settings of Blend File: Barbershop, Compute: CPU-Only. EPYC 9845 - acpi-cpufreq performance was the fastest.

For those after high core count processors for render farms, the EPYC 9845 upgrade from EPYC 97554 was huge for being a one-generation transition and also sticking to the same DDR5-4800 server memory. The power profile / AMD P-State tuning also netted some nice efficiency improvements here too without any measurable cost to the Blender 3D render performance.

OSPRay Studio benchmark with settings of Camera: 1, Resolution: 4K, Samples Per Pixel: 16, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.
OSPRay Studio benchmark with settings of Camera: 1, Resolution: 4K, Samples Per Pixel: 16, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.
OSPRay Studio benchmark with settings of Camera: 1, Resolution: 4K, Samples Per Pixel: 16, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.
OSPRay Studio benchmark with settings of Camera: 3, Resolution: 4K, Samples Per Pixel: 32, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.
OSPRay Studio benchmark with settings of Camera: 3, Resolution: 4K, Samples Per Pixel: 32, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.
OSPRay Studio benchmark with settings of Camera: 3, Resolution: 4K, Samples Per Pixel: 32, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - acpi-cpufreq performance was the fastest.

Intel's OSPRay Studio software suite is another example of a CPU rendering workload showing the excellent opportunities provided by the EPYC 9845 for better performance and efficiency on existing 400 Watt rated server platforms.

OSPRay Studio benchmark with settings of Camera: 3, Resolution: 4K, Samples Per Pixel: 16, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
OSPRay Studio benchmark with settings of Camera: 3, Resolution: 4K, Samples Per Pixel: 16, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
OSPRay Studio benchmark with settings of Camera: 3, Resolution: 4K, Samples Per Pixel: 16, Renderer: Path Tracer, Acceleration: CPU. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
Embree benchmark with settings of Binary: Pathtracer ISPC, Model: Asian Dragon. EPYC 9845 was the fastest.
Embree benchmark with settings of Binary: Pathtracer ISPC, Model: Asian Dragon. EPYC 9845 was the fastest.
Embree benchmark with settings of Binary: Pathtracer ISPC, Model: Asian Dragon. EPYC 9845 was the fastest.

OSPRay Studio along with other Intel oneAPI rendering toolkit software like Embree is well tuned for AVX-512 and thus very much enjoying the full 512-bit data path available with the Zen 5 cores.

LULESH benchmark with settings of . EPYC 9845 was the fastest.
LULESH benchmark with settings of . EPYC 9845 was the fastest.
LULESH benchmark with settings of . EPYC 9845 was the fastest.
OpenRadioss benchmark with settings of Model: INIVOL and Fluid Structure Interaction Drop Container. EPYC 9845 was the fastest.
OpenRadioss benchmark with settings of Model: INIVOL and Fluid Structure Interaction Drop Container. EPYC 9845 was the fastest.
OpenRadioss benchmark with settings of Model: INIVOL and Fluid Structure Interaction Drop Container. EPYC 9845 was the fastest.
ASKAP benchmark with settings of Test: tConvolve OpenMP, Gridding. EPYC 9845 was the fastest.
ASKAP benchmark with settings of Test: tConvolve OpenMP, Degridding. EPYC 9845 - acpi-cpufreq performance was the fastest.
Graph500 benchmark with settings of Scale: 26. EPYC 9845 - acpi-cpufreq performance was the fastest.
NAMD benchmark with settings of Input: ATPase with 327,506 Atoms. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
NAMD benchmark with settings of Input: STMV with 1,066,628 Atoms. EPYC 9845 - acpi-cpufreq performance was the fastest.

From Altair OpenRadioss to NAMD and other HPC software packages, the EPYC 9754 to EPYC 9845 upgrade was very compelling to performance and efficiency. These gains were rather wild from a simple CPU swap and not needing to worry about increasing cooling/power capacity as well as sticking to the existing DDR5-4800 memory. But if planning for heavy production work, the DDR5-6000/6400 memory can be very beneficial for high performance computing workloads to enjoy even greater uplift.

OpenFOAM benchmark with settings of Input: drivaerFastback, Medium Mesh Size, Execution Time. EPYC 9845 - acpi-cpufreq performance was the fastest.
OpenFOAM benchmark with settings of Input: drivaerFastback, Medium Mesh Size, Execution Time. EPYC 9845 - acpi-cpufreq performance was the fastest.
OpenFOAM benchmark with settings of Input: drivaerFastback, Medium Mesh Size, Execution Time. EPYC 9845 - acpi-cpufreq performance was the fastest.
OpenFOAM benchmark with settings of Input: drivaerFastback, Large Mesh Size, Execution Time. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
OpenFOAM benchmark with settings of Input: drivaerFastback, Large Mesh Size, Execution Time. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
OpenFOAM benchmark with settings of Input: drivaerFastback, Large Mesh Size, Execution Time. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.

With roughly the same CPU power consumption, the OpenFOAM CFD software was enjoying much quicker execution times out of the EPYC 9845.

John The Ripper benchmark with settings of Test: WPA PSK. EPYC 9845 - acpi-cpufreq performance was the fastest.
John The Ripper benchmark with settings of Test: WPA PSK. EPYC 9845 - acpi-cpufreq performance was the fastest.
John The Ripper benchmark with settings of Test: WPA PSK. EPYC 9845 - acpi-cpufreq performance was the fastest.
nginx benchmark with settings of Connections: 500. EPYC 9845 was the fastest.
nginx benchmark with settings of Connections: 500. EPYC 9845 was the fastest.
nginx benchmark with settings of Connections: 500. EPYC 9845 was the fastest.

There were very nice gains out of the EPYC 9845 upgrade that extend well beyond the 25% core/thread count advantage over the EPYC 9754.

PostgreSQL benchmark with settings of Scaling Factor: 100, Clients: 800, Mode: Read Write. EPYC 9845 was the fastest.
PostgreSQL benchmark with settings of Scaling Factor: 100, Clients: 800, Mode: Read Write, Average Latency. EPYC 9845 was the fastest.
PostgreSQL benchmark with settings of Scaling Factor: 100, Clients: 800, Mode: Read Only. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.
PostgreSQL benchmark with settings of Scaling Factor: 100, Clients: 800, Mode: Read Only, Average Latency. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.
MariaDB benchmark with settings of Test: oltp_write_only, Threads: 512. EPYC 9845 was the fastest.
MariaDB benchmark with settings of Test: oltp_write_only, Threads: 512. EPYC 9845 was the fastest.
MariaDB benchmark with settings of Test: oltp_write_only, Threads: 512. EPYC 9845 was the fastest.
MariaDB benchmark with settings of Test: oltp_update_index, Threads: 512. EPYC 9845 was the fastest.
MariaDB benchmark with settings of Test: oltp_update_index, Threads: 512. EPYC 9845 was the fastest.
MariaDB benchmark with settings of Test: oltp_update_index, Threads: 512. EPYC 9845 was the fastest.

The AMD EPYC 9845 proved to be a very capable upgrade as well for database servers.

Apache Cassandra benchmark with settings of Test: Writes. EPYC 9845 was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. EPYC 9845 was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. EPYC 9845 was the fastest.
ClickHouse benchmark with settings of 100M Rows Hits Dataset, Third Run. EPYC 9845 was the fastest.

The AMD P-State Powersave + Balanced Memory can also help in achieving even greater efficiency on Zen 5 servers.

OpenSSL benchmark with settings of Algorithm: AES-128-GCM. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
OpenSSL benchmark with settings of Algorithm: AES-128-GCM. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
OpenSSL benchmark with settings of Algorithm: AES-128-GCM. EPYC 9845 - amd-pstate powersave, EPP performance, Balanced Mem was the fastest.
OpenSSL benchmark with settings of Algorithm: AES-256-GCM. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.
OpenSSL benchmark with settings of Algorithm: AES-256-GCM. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.
OpenSSL benchmark with settings of Algorithm: AES-256-GCM. EPYC 9845 - amd-pstate powersave, EPP power, Balanced Mem was the fastest.

For those concerned about ensuring optimal energy efficiency without significant sacrifices to performance, the Balanced Memory Power Profile with AMD P-State power tuning continues to prove to be quite effective.

OpenSSL benchmark with settings of Algorithm: ChaCha20-Poly1305. EPYC 9845 - acpi-cpufreq performance was the fastest.
OpenSSL benchmark with settings of Algorithm: ChaCha20-Poly1305. EPYC 9845 - acpi-cpufreq performance was the fastest.
OpenSSL benchmark with settings of Algorithm: ChaCha20-Poly1305. EPYC 9845 - acpi-cpufreq performance was the fastest.
Kvazaar benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Medium. EPYC 9845 was the fastest.
Kvazaar benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Medium. EPYC 9845 was the fastest.
Kvazaar benchmark with settings of Video Input: Bosphorus 4K, Video Preset: Medium. EPYC 9845 was the fastest.

The AMD P-State tunables as well as the Power Profile (via AMD HSMP) can all be easily managed at run-time if wanting to dynamically cater to changing compute needs.

GraphicsMagick benchmark with settings of Operation: HWB Color Space. EPYC 9845 was the fastest.
GraphicsMagick benchmark with settings of Operation: HWB Color Space. EPYC 9845 was the fastest.
GraphicsMagick benchmark with settings of Operation: HWB Color Space. EPYC 9845 was the fastest.
GraphicsMagick benchmark with settings of Operation: Noise-Gaussian. EPYC 9845 was the fastest.
GraphicsMagick benchmark with settings of Operation: Noise-Gaussian. EPYC 9845 was the fastest.
GraphicsMagick benchmark with settings of Operation: Noise-Gaussian. EPYC 9845 was the fastest.

With only a BIOS upgrade and changing the processor, the EPYC 9754 to EPYC 9845 upgrade proved to be an extremely compelling upgrade without any increased power demands and being compatible with the wide-range of servers/motherboards having a 400 Watt TDP limit.

CPU Power Consumption Monitoring Overview benchmark with settings of Accumulated CPU Power Consumption Monitoring.

Across all the 100+ benchmarks run, the EPYC 9845 had a 234 Watt average compared to 243 Watts with the EPYC 9754 Bergamo. The EPYC 9845 at its default TDP on this AMD Titanite server had a peak consumption of 345 Watts, compared to 398 Watts with the EPYC 9754. For those wanting great performance but concerned about maximizing efficiency, the AMD P-State EPP and Power Profile tuning led to 6.5% lower CPU power use on the EPYC 9845 1P.

CPU Temp Monitoring Overview benchmark with settings of Accumulated CPU Temp Monitoring.

The EPYC 9845 was running slightly cooler on this AMD EPYC 2U server upgrade than with the EPYC 9754, as expected.

Geometric Mean Of All Test Results benchmark with settings of Result Composite, AMD EPYC 9845 Benchmarks vs. EPYC 9754 Linux Comparison. EPYC 9845 was the fastest.

When taking the geometric mean of all the completed benchmarks, the upgrade from the EPYC 9754 to EPYC 9845 delivered 1.36x the performance. Very nice uplift with the same to slightly lower power consumption and being suitable for use in existing AMD Socket SP5 motherboards having a 400 Watt power limit. That's also while still using DDR5-4800 memory. As shown in prior Turin with DDR5-4800 vs. DDR5-6000 memory benchmarks, there can be very sizable gains with the faster memory -- all the more so too for the EPYC 9845 given the increased number of cores contending for the twelve memory channels. Simply upgrading the BIOS and swapping the CPU without having to resort to a memory upgrade and any increased power/cooling requirements while achieving 1.36x the performance is a very persuasive generational upgrade. Plus with the AMD EPYC 9005 server processors is the AMD P-State support for allowing greater energy efficiency if desired.

Those wanting to go through all 100+ benchmarks in full can find them via this result page.

Read More

Some Chromecasts are giving ‘Untrusted device’ errors today

Reports have been cropping up that second-generation Chromecasts and music-streaming Chromecast Audio devices are failing, displaying an “Untrusted device” error on the device they’re trying to cast from, reports 9to5Google. Text from screenshots of the error posted to Reddit says the device “couldn’t be verified. This could be caused by outdated device firmware,” leading some […]

Read More

[杂谈]新号定级真逆天啊！

看佛珠打500强局，对面elo直接给个新号76，全程输出4000出头，路都走不利索那种，就是那种初始头像，默认配色的76

还有一把，对面T白金，被这边500强T按着锤

然后自己elo了，给了个1000伤害新手托比昂，打不出来就换了个默认配色的76

不是亲眼看见很难相信这是500强局的匹配

Read More

网友喊话海尔希望推出“懒人洗衣机”，官方称已让工程师连夜开发

IT之家 3 月 10 日消息，昨日有网友在海尔集团公司官方抖音评论区留言，向海尔集团董事局主席、首席执行官周云杰“许愿”—— 一款能同时洗外衣、内衣、鞋子、袜子的“懒人洗衣机”。他表示：“周总好，要是你们研发生产这个，我攒钱买。”如图所示，这款洗衣机配备了 3 个滚筒，分别用于洗涤外衣、内衣、袜子，下方还有专门用于洗鞋的区域。

海尔官方回应称，周云杰已经看到了这个设计，并让工程师连夜开发中，最后还“期待”这名网友成为该新品的代言人。

海尔 2024 年全球营收 4016 亿元，同比增长 8%；全球利润总额 302 亿元，同比增长 13%。集团子公司海尔智家连续 7 年入围《财富》世界 500 强榜单。

据IT之家此前报道，周云杰在今年年初表示，海尔要成为全球市值最大的家生态公司，同时成为中国一流的智能制造标杆，“未来是液态的时代，需要流动的组织来支撑；未来是数据的时代，数据是最重要的资产；未来是使用权的时代，使用权优于所有权；未来是人工智能的时代，要么与 AI 同进化，要么被 AI 淘汰。”

Read More