arXiv subjects

Chufeng Chen

Publications and source records attributed to Chufeng Chen.

5 recordsLinked to original sources

arXiv2026

SieveIVF: Threshold-Aware IVF Execution for Large-Scale Training Data Deduplication

Embedding-based training data deduplication retrieves candidate duplicate edges above an application similarity threshold, but fixed-probe inverted-file (IVF) search ignores this predicate when giving every query the same partition budget. Across four Hunyuan workloads, qualifying neighbors appear early despite sharply varying search depths. We present SieveIVF, a threshold-aware IVF executor that stops after $W$ consecutive searches find no qualifying candidate. The systems challenge is to preserve partition-major batching when each query's remaining work depends on prior results. Continuous batching groups ready queries by partition. A lookahead scheduler layers on top, exposing only work committed by the stopping rule to increase concurrency without changing stopping decisions or returned results. We implement SieveIVF in Lance. At $W=8$, SieveIVF is $4.1$--$7.6\times$ faster than fixed-probe IVF on four 10M Hunyuan workloads and $6.1$--$8.4\times$ faster on two public 100M workloads under the same index and search parameters, with pooled filtered top-10 recall losses of $0.03$--$1.13$ percentage points on Hunyuan and $1.43$--$2.29$ percentage points on the public workloads. These results show how an application predicate can guide IVF work allocation without changing the index or bounded top-$k$ interface.

cs.DB↗

arXiv2020

1.28 and 5.12 Gbps multi-channel twinax cable receiver ASICs for the ATLAS Inner Tracker Pixel Detector Upgrade

We present two prototypes of a gigabit transceiver ASIC, GBCR1 and GBCR2, both designed in a 65-nm CMOS technology for the ATLAS Inner Tracker Pixel Detector readout upgrade. The first prototype, GBCR1, has four upstream receiver channels and one downstream transmitter channel with pre-emphasis. Each upstream channel receives the data at 5.12 Gbps through a 5 meter AWG34 Twinax cable from an ASIC driver located on the pixel module and restores the signal from the high frequency loss due to the low mass cable. The signal is retimed by a recovered clock before it is sent to the optical transmitter VTRx+. The downstream driver is designed to transmit the 2.56 Gbps signal from lpGBT to the electronics on the pixel module over the same cable. The peak-peak jitter (throughout the paper jitter is always peak-peak unless specified) of the restored signal is 35.4 ps at the output of GBCR1, and 138 ps for the downstream channel at the cable ends. GBCR1 consumes 318 mW and is tested. The second prototype, GBCR2, has seven upstream channels and two downstream channels. Each upstream channel works at 1.28 Gbps to recover the data directly from the RD53B ASIC through a 1 meter custom FLEX cable followed by a 6 meter AWG34 Twinax cable. The equalized signal of each upstream channel is retimed by an input 1.28 GHz phase programmable clock. Compared with the signal at the FLEX input, the additional jitter of the equalized signal is about 80 ps when the retiming logic is o . When the retiming logic is on, the jitter is 50 ps at GBCR2 output, assuming the 1.28 GHz retiming clock is from lpGBT. The downstream is designed to transmit the 160 Mbps signal from lpGBT through the same cable connection to RD53B and the jitter is about 157 ps at the cable ends. GBCR2 consumes about 150 mW when the retiming logic is on. This design was submitted in November 2019.

physics.ins-det↗

arXiv2016

Detailed study of the column-based priority logic readout of Topmetal-II- CMOS pixel direct charge sensor

We present the detailed study of the digital readout of Topmetal-II- CMOS pixel direct charge sensor. Topmetal-II- is an integrated sensor with an array of 72X72 pixels each capable of directly collecting external charge through exposed metal electrodes in the topmost metal layer. In addition to the time-shared multiplexing readout of the analog output from Charge Sensitive Amplifiers in each pixel, hits are also generated through comparators with individually DAC settable thresholds in each pixel. The hits are read out via a column-based priority logic structure, retaining both hit location and time information. The in-array column-based priority logic is fully combinational hence there is no clock distributed in the pixel array. Sequential logic and clock are placed on the peripheral of the array. We studied the detailed working behavior and performance of this readout, and demonstrated its potential in imaging applications.

physics.ins-det↗

arXiv2016

Test of \textit{Topmetal-${II}^-$} In Liquid Nitrogen For Cryogenic Temperature TPCs

\textit{Topmetal-${II}^-$} is a highly pixelated direct charge sensor that contains a 72${\times}$72 pixel array of 83$μ$m pitch size. The key feature of \textit{Topmetal-${II}^-$} is that it can directly collect charges via metal nodes of each pixel to form two-dimensional images of charge cloud distributions. \textit{Topmetal-${II}^-$} was proved to measure charged particles without amplification at room temperature. To measure its performance at cryogenic temperature, a \textit{Topmetal-${II}^-$} sensor is embedded into a liquid nitrogen dewar. The results presented in this paper show that \textit{Topmetal-${II}^-$} can also operate well at this low temperature with a noise (ENC) of 12 e$^-$ lower than that at room temperature (13 e$^-$). From the noise perspective, \textit{Topmetal-${II}^-$} is a promising candidate for the next generation readout of liquid argon and xenon Time Projection Chamber (TPC) used in experiments searching for neutrinoless double beta decay and dark matter.

physics.ins-det↗

arXiv2015

A Low-Noise CMOS Pixel Direct Charge Sensor, Topmetal-II-

We report the design and characterization of a CMOS pixel direct charge sensor, Topmetal-II-, fabricated in a standard 0.35um CMOS Integrated Circuit process. The sensor utilizes exposed metal patches on top of each pixel to directly collect charge. Each pixel contains a low-noise charge-sensitive preamplifier to establish the analog signal and a discriminator with tunable threshold to generate hits. The analog signal from each pixel is accessible through time-shared multiplexing over the entire array. Hits are read out digitally through a column-based priority logic structure. Tests show that the sensor achieved a <15e- analog noise and a 200e- minimum threshold for digital readout per pixel. The sensor is capable of detecting both electrons and ions drifting in gas. These characteristics enable its use as the charge readout device in future Time Projection Chambers without gaseous gain mechanism, which has unique advantages in low background and low rate-density experiments.

physics.ins-det↗